Datasets:

mimiklee

/

masterthesis-longt5-sum-8k

like 0

INTRODUCTION A person's internal sense of their gender, which may or may not match the gender that was given to them at birth based on their biological sex, is referred to as gender identity. While gender is a social and cultural construct that includes the expectations, roles, and behaviors associated with being male, female, or another gender identity, sex refers to the biological characteristics of a person's reproductive system. Many people identify as either male or female, regardless of the sex assigned to them at birth. Some people, however, may identify as transgender, non-binary, gender-queer, or another gender identity that does not fit into the conventional binary categories of male and female if their gender identity does not match their assigned sex. --- 4th INCLAR Gender is a cultural notion that aims to distinguish between men and women who succeed in society in terms of roles, behavior, mindset, and emotional traits, according to the Women's Studies Encyclopedia. [1] A crucial component of a person's entire identity, gender identity is acknowledged as a protected feature in many nations, including Indonesia. For a society to be more inclusive and egalitarian for everyone, regardless of gender identity, it is imperative to recognize and respect people's gender identities. In contrast to the preceding practice, the measuring instruments designed to evaluate femininity and masculinity only included psychological qualities derived from gender stereotypes in their list of items. The personality qualities chosen for these tests were therefore more stereotypical of one sex than the other. [2] According to Article 30 of Law Number 39 of 1999 on Human Rights (Human Rights Law), everyone has the right to feel safe and secure and to be safeguarded from threats and terror. The Republic of Indonesia's Constitution, namely the Constitution of The Republic of Indonesia 1945 (1945 Constitution of the Republic of Indonesia), the Human Rights Law, the Universal Declaration of Human Rights (UDHR), and other laws protect the right to experience this security. [3] According to Joel Feinberg ( Joel Feinberg: 2004), Human rights are moral rights pertaining to anything fundamentally significant that are unconditionally, inviolably owned by all civilizations. Whether these rights fall into the category of "moral" in the strict sense, I leave them as open issues that must be resolved through argument and not definitions." [4] Any behavior or action intended to frighten, intimidate or annoy another person is called harassment. This can occur in a variety of contexts, including the workplace, school, online media, or public places, and can take many forms, such as physical, verbal, psychological, or sexual. Here are some instances of harassment: (a) verbal abuse: this might involve making disparaging remarks, hurling insults, or threatening someone because of their gender, race, ethnicity, religion, or sexual orientation; (b) sexual harassment: can take the form of unwanted advances, remarks, actions, or physical contact that is suggestive of sex; (c) physical violence, assault, or unwelcome touching can all be considered forms of physical harassment; (d) cyber harassment: cyberstalking, online bullying, and the dissemination of untrue information via social media; (e) psychological harassment: threats, intimidation, or other actions that are meant to instill fear or anxiety in another person. DOI --- METHODOLOGY/ MATERIALS Numerous data-generation techniques are used in empirical legal studies. Depending on the situation at hand, researchers can use random sampling in a variety of ways. Researchers can compare their overlap through data analysis. Empirical legal research seeks to establish facts regarding the unknowable. The presentation of empirical legal research findings is the final step, and it requires documentation. [5] Primary study data were collected directly from LGBTQ+ people through surveys sent at random on social media, and the data were then evaluated using descriptive analysis. The term "primary data" refers to information or data that researchers have directly accessed from a source. [6] --- RESULTS AND DISCUSSIONS Based on questionnaires that have been distributed through social media to 41 respondents consisting of 68.3% women and 31.7% men and spread across several cities in Indonesia, including Jakarta, Depok, Tangerang, Bekasi, Medan, and Bali, 82.9% expressed disapproval of the existence of LGBTQ+ people. It's interesting to note that 85.4% of respondents believe that the existence of LGBTQ+ individuals is a religious offense, while the remaining 14.6% believe that this is a human right and a psychological condition. The respondents' statements of disapproval were stated because the existence of LGBT people was a violation of religious rules, while their statements of agreement were considered a human right. Same-sex relationships are illegal in many states, and homosexual, bisexual, and transgender people are viewed as immoral. As a result, discriminatory acts and violence motivated by homophobia or transphobia are grave problems in which the fundamental rights of people are disregarded, leading to the stigmatization of the group and, ultimately, to other violations of their human rights. LGBT in Indonesia is still a taboo subject, especially for groups whose thoughts are based on religion. Most of them criticized the behavior and sexual orientation of the LGBT group. MUI has even issued a fatwa against the practice of intercourse unions and same-sex marriages. [7] DOI 10.18502/kss.v8i21.14706 4th INCLAR Indonesia has no national laws against homosexuality except for Law (UU) No. 1 of 1974 concerning Marriage. This law stipulates that only marriages between heterosexual couples are valid. [8] According to Article 1 of the Constitution, marriage is the union of a man and a woman as husband and wife in order to create a joyful, eternal family (home) founded on faith in the One Supreme God. The WHO 1990 removed homosexuality from its list of mental disorders, recognizing that sexual orientation is not a disease or a disorder. Additionally, the WHO has acknowledged the unique health challenges faced by LGBTQ+ individuals, including discrimination and violence, and has called for efforts to address these issues. The House of Representatives has no reluctance to purge public spaces of LGBTrelated materials, despite the fact that this law has arguably been the most divisive and has been the subject of debates for more than four years. According to The Jakarta Post, " LGBT behavior" is one of the 12 conditions for broadcasting content to be forbidden, and is specifically mentioned in article 61 of the draft amendment. Additionally, Article 140 mandates that all films, dramas, and ads be vetted by the censorship office to remove any content linked to the LGBT community. Legislators concurred that "LGBT behavior" is harmful to children and goes against "Indonesian culture" despite the vague description of what constituted it. [9] The goal of the May 17 celebration of the International Day Against Homophobia, Biphobia, and Transphobia is to coordinate global efforts to promote LGBT rights activities and increase public awareness of breaches of LGBT rights. Over 130 nations had observed the commemorations as of 2016. The IDAHO Committee was founded by the original organizers of the May 17 event, which was first known as the International Day Against Homophobia, in order to coordinate grassroots activities taking place in many nations, publicize the day, and push for official recognition. The World Health Organization (WHO) chose that day to honor its 1990 decision to remove homosexuality from its list of international diseases. [10] Discrimination and prejudice towards LGBTQ+ persons are damaging and unjustified, and they should not be tolerated. The existence of LGBTQ+ people throughout history and cultures must be acknowledged, and their existence is a natural and normal variation of human identity and expression. Disapproval of LGBTQ+ people's presence can take many forms, including denial of their fundamental rights, discrimination in jobs, and health care, or even violent or hateful acts. The mental and physical health of LGBTQ+ people, as well as those in their families and communities, may suffer significantly as a result. Another survey among trans-Australian adults, a large community-based survey with 928 participants reported persistent, alarming health statistics, including high rates of self-reported mental health morbidities like anxiety and depression as well as selfreported self-harm (63%) and attempted suicide (43%). Discrimination was frequently encountered, especially in hospital settings (26%). In addition, most individuals (63%) reported verbal abuse; fewer (22%) reported physical attacks due to their trans identity. [11] One of the rejections of LGBTQ+ people is verbal abuse. Emotional abuse includes verbal abuse. It occurs when someone verbally abuses, dominates, mocks, manipulates, or denigrates another person, detrimental to that person's mental health. [12] Verbal abuse, such as insults, threats, and insulting remarks, can have a very detrimental impact on LGBTQ+ people's mental and emotional health as well as contributing to a culture of fear and isolation. It is critical to acknowledge that LGBTQ+ persons are entitled to the same protections and human rights as everyone else. The impact of verbal abuse is immeasurable. The shock and trauma are often disabling. Millions of people suffer from parents, peers, and even cultures that define them. Numerous men and women have also shared with me their experiences of suffering for years as a result of what others have said about them. [13] Many LGBT persons also encounter verbal harassment, which is comparable to catcalling and includes insulting comments, whistling, and sudden groping. Catcalling is the practice of yelling annoying, insulting, and frequently sexually suggestive remarks at someone in public. [14] Due to the fact that catcalling occurs spontaneously, it is frequently missed as an act that is impermissible and falls under the penal code. [15] Promoting acceptance and respect for LGBTQ+ persons and their identities is crucial, as is working to build an inclusive culture that supports everyone, no matter what their gender, sexual orientation, or preferred manner of expression may be. This can be accomplished by spreading awareness, advocating for change, and supporting legislation and regulations that uphold the rights and welfare of LGBTQ+ people. In Indonesia, speaking about LGBTQ+ persons is generally frowned upon or odd, hence verbal abuse is still common. Unfortunately, there is still significant social stigma and discrimination against LGBTQ+ people in Indonesia, which can result in verbal abuse and other types of mistreatment. There have been claims of rising hostility toward the LGBTQ+ population in Indonesia in recent years, including the adoption of legislation that makes homosexuality and transsexual identities illegal. In Indonesia, advocacy organizations and individuals are working to advance greater acceptance and comprehension of the LGBTQ+ population and to press for societal DOI 10.18502/kss.v8i21.14706 4th INCLAR and legislative changes that will protect their rights. The underlying preconceptions and stereotypes that fuel discrimination against LGBTQ+ persons in Indonesia can be addressed through education, awareness campaigns, and community outreach. Demolishing these destructive notions requires a collective effort from individuals, communities, and institutions, and it requires ongoing commitment and vigilance to ensure that progress is sustained over time. By working together to promote greater acceptance and understanding of LGBTQ+ individuals, we can help create a more inclusive and equitable society that values and celebrates diversity in all its forms. --- CONCLUSION AND RECOMMENDATION Due to possible social stigma, discrimination, and exclusion, LGBTQ+ people may be more likely to experience psychological problems. Negative mental health effects might result from the strain of adjusting to a society that does not fully embrace or comprehend one's sexual orientation or gender identity. Understanding that these psychological issues are not caused by a person's sexual orientation or gender identity is vital. Instead, these issues are caused by the stress and discrimination that they may experience as a result of how society perceives them instead. To help LGBTQ+ people deal with these difficulties and to advance greater acceptance and understanding of LGBTQ+ identities in society, it is crucial to offer them resources and support. Historically, the majority of societies have condemned homosexual, bisexual, and transgender people, viewing them as immoral or as having a mental or physical disease. Many societies also punished such individuals. As well as criminalizing same-sex relationships, many jurisdictions also view homosexual, bisexual, and transgender people as morally repugnant. Therefore, homophobic or transphobic discriminatory acts and violence are a grave problem in which basic human rights are disregarded. Privatization of essential rights, such as those to family life, privacy, and social security, encourages stigmatization of the community and, as a result, further violations of their human rights. LGBTQ+ people frequently experience family isolation, verbal or physical harassment, corporal punishment, name-calling bullying, as well as physical assault. Parents frequently make such children endure various reparative therapies to change their sexual orientation and gender identity, which further equates to child abuse, in order to suit those within the heterosexual population or the people whose gender identity and gender role conform to anatomic sex. All of these violent acts and discriminatory practices have a negative impact on the mental and physical health of the LGBTQ+ community and raise worrisome rates of drug misuse, HIV, and other STDs. Although the problem is serious, there hasn't been much attention given to it. Verbal abuse directed towards LGBTQ+ people can have a serious negative impact on their mental and emotional health. It can result in mental health problems like anxiety and depression as well as feelings of guilt, self-doubt, and low self-esteem. Verbal abuse can also foster an environment that is unfriendly and hazardous for LGBTQ+ people and can help to foster a culture of fear and exclusion. In addition to its impact on individual well-being, verbal abuse against LGBTQ+ individuals can also contribute to broader social issues, including a lack of acceptance and understanding of LGBTQ+ identities, and a failure to recognize the basic human rights and protections to which all individuals are entitled. Stopping the spread of false information about some discriminatory situations, like homosexuality, and providing factual education are important components of campaigns for social equality in society. Trying to understand the main idea beyond the volume of murky material frequently disseminated is one of the hardest obstacles in helping society understand LGBTQ+. Disposing of misconceptions, myths, prejudices, and lies is critical to have a productive conversation on LGBTQ+ problems.

Verbal abuse toward individuals who identify as LGBTQ+ are often rooted in rigid stereotypes and societal norms surrounding gender and identity. This type of abuse can range from derogatory comments and slurs to threats and harassment and can significantly impact the mental health and well-being of those who experience it. This study aimed to focus on verbal rejection, which is indirectly and unwittingly, carried out in the immediate environment of LGBTQ+ people. The main method of research conducted was primary data obtained directly from LGBTQ+ people, questionnaires were circulated randomly through social media, and then the data were analyzed by descriptive analysis. According to the study's findings, verbal violence against LGBTQ+ individuals frequently occurs in Indonesia, since it is still considered to be a taboo or odd. In order to establish a more inclusive and accepting society for everyone, regardless of a person's gender identity or sexual orientation, it is imperative to confront and destroy these harmful presumptions.

Introduction Demographic change in Europe is increasingly becoming a major policy challenge. Predominantly rural regions currently account for 28% of Europe's population. However, over the past decades, a shrinking population has become the normal trajectory for many rural regions as agriculture has been restructured and population and employment have become increasingly concentrated in urban centers (Espon, 2017). By 2050, the population of Europe's urban regions is projected to increase by 24.1 million persons and will provide a home to almost half of the EU-28 population (Eurostat, 2016). By contrast, the population of predominantly rural regions is projected to fall by 7.9 million. Recently, rural areas have been affected by the phenomenon of village shrinkage (Espon, 2017). Large-scale urbanization has been happening autonomously (IRMiR, 2021); it has been actively advocated for by individuals, collective powers, and states since the 19th century (Dudek et al., 2016). Cities are now the main centers for technology development, innovation, education, commerce, administration, transportation, medical care, human resources, etc. Further development of cities is discussed within the idea of smart cities, however, the smartness as such also refers to the rural areas (Gerli et al., 2022). At the same time, large-scale urbanization changed the lifestyle of people and very often, individuals with an urban lifestyle are choosing to live outside of cities. For this, they often choose green and ecologically clean areas. Living in rural areas, they maintain employment in urban companies and organizations and continue to live in accordance with the urban lifestyle (Gawro<unk>ska-Nowak, 2021). According to the definition presented by European Network for Rural Development (Cork 1.0, 1997;ADE et al., 2023), smart villages are communities in rural areas that apply innovative solutions to improve their resilience, taking advantage of local strengths and opportunities. They are based on a participatory approach to develop and implement their strategy to improve their economic, social, or environmental conditions, in particular mobilizing solutions offered by digital technologies. Smart villages benefit from cooperation and alliances with other communities and actors in rural and urban areas. The initiation and implementation of smart village strategies are based on existing initiatives and can be funded by a variety of public and private sources (ENRD, 2023;ADE et al., 2023). An important role in the implementation of the concept of smart villages refers to solutions to improve the quality of life, underinvestment, ageing society, village shrinkage, increasing the quality of services and safety, respect for the local environment, insufficient job opportunities, and digital divide. What can be underlined in particular is the importance of not associating this concept with issues such as digitization of villages, providing ready-made solutions to existing social problems related to the above challenges, and repetition of existing solutions (Boba, 2020;ADE et al., 2023). The true foundation of a smart village is (not just the territorial community of people registered in a given village, but also) the local community that lives in the village and that identifies with its place of residence and that acts in agreement with neighbors. Another key aspect is trust in the local (village) government and local leaders that they care for the common good of the place and thus (in small steps) improve the quality of life in the village which, thanks to cooperation, step by step, becomes a smart village. The aim of the smart village concept is to focus (Cork 2.0, 2016) on energy, vision, and commitment of local people towards local action. The examples of projects and initiatives identified so far clearly show that smart villages start with local people organizing themselves around a common problem or shared vision to implement some form of 'action plan' to achieve a specific goal (Cork 2.0, 2016). Ideas of smart villages refer in concept to the ideas of smart cities. The question under discussion is how these two dimensions are interconnected (Kusio, 2022, Table 1). The objective of this paper is to show the role of local people partnership and partners from urban centers in local development concepts of rural areas in order to find the potential for future cooperation. --- Literature Review Some researchers and proponents of urbanization argue that more than 99% of humans might live in cities by the end of this century and that the valuing and planning of rural areas should focus on urban needs (Adamashvili et al., 2020;Sardaro et al., 2021). Some even regard rural areas as places that are suitable exclusively for growing crops and keeping livestock (Daniels, 2001;European Commission, 2017;Guzal-Dec, 2018). Alternatively, it can be proposed that the urban lifestyle will replace the rural lifestyle, especially in the case of traditionally small villages located within the influence of large metropolises, and will become widespread in all forms of urban and small rural human settlements. Residents of urban places, in search of a healthy lifestyle and green and less urbanized areas, will choose villages as their places of residence; villages, where living conditions will be suitable enough to lead an urban lifestyle. The concept of smart villages, in the European Union, arises in reference to cohesion and strengthening rural development, while in other highly developed countries, intelligent rural development is related to spatial planning and relies on actions that suppress the spontaneous growth of cities (urban sprawl) (Daniels, 2001). Within the European Union documentary (European Commission, 2017), the smart villages concept refers to rural areas and communities that want to base their development on their strengths and resources. Traditional and new networks and services in smart villages are strengthened by the means of digital technologies, telecommunications, innovation, and better use of knowledge, for the benefit of residents and enterprises (Guzal-Dec, 2018). Smart villages are communities located in rural areas using innovative solutions in different spheres of economic activity to improve their socio-economic well-being, building on local assets and opportunities (Boba, 2020). Smart villages is also a term used to describe a new concept in EU policymaking based on the functioning of traditional and new networks and services enhanced by digital, telecommunication, innovation, and better use of knowledge, for the benefit of citizens and businesses (Boba, 2020). The genesis of the smart villages concept should be associated in particular with the signing of the Cork 2.0 Declaration in 2016 (Cork 2.0, 2016), the main objective of which was to answer the question: What may be done in order to make the inhabitants of rural areas live better against the background of the ongoing structural changes in the economies of countries and regions? Solving such a problem requires focusing primarily on the following questions (Boba, 2020): a) How to effectively combine financial resources from different funds? b) How to implement different actions and exploit emerging synergies? c) How to use technology to stimulate rural development? Solutions to the above questions have started to be developed in 2019-2020 as part of the concept formulation at the level of EU member states. The EU member states are encouraged to develop strategies and interventions as a part of the implementation of the smart villages concept, presenting proposed answers and solutions to the problems present in rural areas. In this kind of strategies (Kalinowski et al., 2021), among the specific features and determinants of the development of rural areas, which drive the implementation of the smart villages concept (Guzal-Dec, 2018), the ones worth mentioning are (Foray, et al., 2009;Da Rosa Pires et al., 2014;Teräs et al., 2015): a) the spatial distance, relatively less developed transport and communication network, lack of the network cooperation, scientific and research organizations and institutions, limitations in building a creative economy; b) insufficient number of enterprises and farms, including entities which are introducing innovations, insufficient number of organizations responsible for intermediation in the exchange of the new knowledge and its adaptation to local conditions; c) difficulties in initiating the innovative projects and raising funds for the development of innovation. --- Methodology Creative problem solving directed research towards heuristic methods; therefore, towards the totality of methods and rules of conduct for making the most appropriate decisions in complex situations that require the analysis of available information, as well as the prediction of future phenomena (Armstrong, 2001). Heuristic methods are based on creative thinking and logical combinations and can be found in the field of logic: it is it is the ability to find new facts and connections between them, as well as to detect truths in this way (Cie<unk>lak, 2008, p. 201). This text presents only selected fragments of very detailed field studies titled: space use and space regeneration of rural settlements, conducted in 2020-2021 by the Szczecin University staff and students of Spatial Management (Pirveli et al., 2021). Therefore, here are discussed only the methods directly related to the content in the article and not all tools used during the field research. The research questions identified in the introductory part of this paper are directly related to village G<unk>dno development and the participation of local and urban partners in this development. For the creation of G<unk>dno's profile, Desk Research was used (Smith, 2008). This method is based on the analysis of the available data sources, including in particular their compilation, mutual verification, and processing (BIOSTAT). The Desk Research method does not have a specific timeframethe acquisition and exploration of data was performed during the entire research project due to the fact that the documents are often being updated and combined with the materials obtained during the remaining stages of the study. Outlining a brief historical profile of the village G<unk>dno was needed to understand the residents' attitude towards the place. A smart village is, above all, the willingness to cooperate and bottom-up building (Nardone et al., 2010;Konecny, 2019;ENRD, 2022) of a development strategy for "one's own place", as well as local and urban-rural partnership. Desk research was used to present a brief historical outline of the village. Because G<unk>dno is a traditionally small village that: a) since the early Middle Ages has always been a small rural place and this place (paraphrasing Lefebvre, 1967) has developed its own law, b) from the mid-20th century to the present day it has undergone three fundamental political, systemic, socio-cultural and economic transformations, which undermined trust in the ruling bureaucracy, c) it was planned and built by residents with different cultural roots than the current residents from villages scattered in the east, who should feel a bond with each other and with the place; it is not easy, especially in the context of the unstable past few decades. It is very important to stress that current residents themselves create the foundations and rules for cooperation and sharing common space. While building the profile of G<unk>dno, the research team conducted in-depth interviews with residents, local NGOs, other entrepreneurs, and visited local and regional governments. The aim of the conversations was to obtain answers from the "other" side about troubling issues (such as changing the route of the road from G<unk>dno to Moryn; the usefulness of the new community center and the gym under the cloud for the residents; what the respondent -a person or a company or an officecan do for G<unk>dno; what infrastructural changes are necessary in the near future; whether there is a need for public transport and whether there is a lack of grocery stores for the village residents; the quality and usage of wifi in the village; the sufficiency and origin of the finances from which numerous local events are financed, etc.). A case study can be an exclusive research method as well as one of the research methods for making observations and indicating regularities in the occurring and observed processes (Yin, 2018). A particular recommendation to use the case study method is the existence of a need for an in-depth description of social phenomenon (Yin, 2018). The presentation and explanation of a social phenomenon, which includes facts and events and has an idiographic approach (Lachiewicz and Matejun, 2010), entitles us, among others, to apply the case study method, which, unlike the theoretical approach, is based mainly on empirical data (Matejun, 2011). --- Results and Discussion The village of G<unk>dno (Guhden -German-language name from the period before World War II) is located on Lake Morzyckie in the Mory<unk> commune, West Pomeranian Voivodeship (GPS 14.421667, 52.855278). G<unk>dno, originally a fishing settlement, was founded in the early Middle Ages. Moryn and G<unk>dno are located on opposite sides in the narrowest part of Lake Morzycko. In the past, fishermen traveled this distance in boats. Currently, there is great potential for launching water transport between the beach in G<unk>dno and the restaurant and hotel in Mory<unk>, which is located in the former fishing harbor. In the year 1464, G<unk>dno was a family seat of the Teutonic family from the castle in Mory<unk>. Since then, this small village has changed several owners, including the von Schöning family from 1608; the von Papstein family from 1765; Friedrich von Mühlheim from 1800 (who, around 1830, built an impressive residence: palace, farm buildings, and a park). At that time, there were also brickworks in G<unk>dno; the von Globig family from Saxony in the second half of the 19th century. The von Rohr family owned it until 1945. During the war, the village did not suffer significant damage; however, just after the war, it was devastated and plundered. The new settlers, four families, found the village plundered and the remains of a once thriving family village. For them, restoring former stability likely meant recreating the past of those whose nation had caused the war devastation of their homeland and displacement from their small homelands in the east. This was more natural for them, even if not necessarily approved by the then socialist authorities, influenced by Soviet ideology, which aimed to build a socialist homeland of people without class differences, emphasizing the domination of workers and kolkhoz farmers. The equivalent of collective farms in these areas were PGR/SAF, i.e., State Agricultural Farms. The first Polish settlers, repatriates from Wolin (villages Czersko and Czerniawka), arrived in the village in 1946; they each received 4 hectares of land from the government. Between 1946-50, a State Agricultural Farm operated in the plundered village with only 4 members of PGR; later, PGR was moved to a neighboring village, leaving a recreation resort (o<unk>rodek wypoczynkowy) for State Agricultural Workers in G<unk>dno. Currently the complex in an attractive location is fenced and devastated; since the 1990s, it has changed several owners; respondents could not indicate who the current owner was. G<unk>dno has never been a large village. According to the National Census of Population and Housing the population in the village in 2011 was 95 people. Nowadays with a total area of 24.7081 ha and 123 inhabitants (status: 2021; 2.9% of the Mory<unk> commune's inhabitants), it is not affected by unemployment. Also, the demographic structure does not indicate a problem with aging or depopulation. Rural demographic dependency ratios are as follows: number of inhabitants -123 = 100%; including: in pre-working age (<unk>18 years) -22.9%, in working age (18-59/64 years) -56.8%, in post-working age (59+/64+) -20.3%. Demographic rates of this village are definitely higher than the corresponding indicators of other villages, or for the voivodeship and Poland as a whole. Village residents also have better housing conditions than residents of other agricultural areas in the region or country. According to 2022 data, the average usable area of a real estate put into use is 146.50 m 2 (much larger for the West Pomeranian Voivodeship and in Poland). Moreover, the village is supplied with water and partly with sewage. The Energy Station in Mieszkowice provides ongoing maintenance of energy infrastructure for village residents. The owners of individual family farms have a closed irrigation system and use rainwater to water their plants. During the period of field research, there was no complete sewage system in the village; some households had septic tanks; however, they did not indicate any problem due to the fact that the village has not been connected to the general drainage system. Everyone was aware that establishing a sewage system was included in the development strategy plan. No major road or railway passes through G<unk>dno. The nearest railway station is in the village of Witnica. G<unk>dno can be reached by bike and car using the local (municipal) road, which connects this place, from the east, to national road No. 31 and on the west (road leading to the state border with Germany) -to provincial road No. 125. The village is within a maximum two-hour drive by car to both Polish large cities (Szczecin, Pozna<unk>) and the European metropolis of Berlin. Hence, the town attracts a sufficient number of visitors who want to spend short or longer holidays here, as well as seasonal workers who, for financial reasons, prefer to rent accommodation in G<unk>dno and work in a company on the German or Polish side (Kostrzyn-S<unk>ubice Special Economic Zone). For visitors, apart from the campsites, there is a thriving private recreation center in G<unk>dno (they have their e-profile on various appropriate websites). It's also possible to rent guest rooms from other villagers. Regarding food, one can cook their own by buying fresh products from farmers, including fish, or use the cafes and restaurants in Mory<unk> (distance is 2 km). On the one hand, the lack of busy communication arteries, and on the other hand, the lack of industrial plants here, ensures G<unk>dno and the entire Moryn commune are ecologically clean; the permissible values for sulfur dioxide, nitrogen dioxide, and PM10 are not exceeded. Only the value of benzo(a)pyrene exceeds the permissible standards (SRGM, 2016(SRGM, -2022)). On the ecologically clean area of this small village, there is 1 building (Palace from 1840) registered in the Register of Monuments of the National Heritage Institute (NID, 2023) and 14 forms of nature protection (1 -Natura 2000 area, 10 -Natural Monuments, 2 -Ecological Utilities, 1 -Nature and Landscape Complex) (CRFOP, 2023). Also, the Association for the Development of G<unk>dno Village operates here. The president of the association is the village head of G<unk>dno (Mrs. Katarzyna Kupczyk); she combines her activity in the village government and the association and cooperates with local and regional partners; as a result, numerous events are organized here for residents and tourists. These include meetings during free time and those resulting from the calendar of local holidays, concerts and performances by invited actors and musicians, cross-border events for adults and children, meetings on the beach, and more importantly -an annual bicycle rally (titled "On the trail of Mr. Samochodzik"), which has become a regional tourist product. The village has an aura of mystery. Together with the manor park, the ruins of a classicist palace from the 18th century, and the forms of protected nature, the mystery of this place is related to legends, scattered and hidden in the village greenery. Some of these legends are associated with specific places in the village -at these points, there are graphically decorated descriptions telling about the enchanted princess Almanda, the Seven Brothers, the mysterious rising of the cross on the church tower, the Clay Farmer, etc. Villagers, visitors, and participants of the mass sports event highly appreciate both the village landscape and the accompanying attractions. Many of them also mention observing the sunset from the lake shore or enjoying the silence and the opportunity to contemplate nature as an advantage of the place. As part of the operational objective "infrastructure" (SRGM, 2016(SRGM, -2022 pp. 37-43) pp. 37-43), it is planned to create a local (urban-rural) functional area Mory<unk> -G<unk>dno -Przyjezierze. This is of great importance for G<unk>dno; it will allow -on the one handto increase the tourist potential and development opportunities for the small village; on the other hand -it will create the opportunity to implement road, pedestrian, and bicycle investments important for residents and also for tourists. Construction, reconstruction, modernization, and renovation of pedestrian, and bicycle investments on the Mory<unk> -G<unk>dno section will improve traffic safety, expand the lake's protection zone, and acquire new recreational areas by the bay. --- Conclusion There is no single "right" model for urban-rural partnerships (Konecny, 2019;Nardone et al., 2010;ENRD, 2022). The EU provides various platforms for information exchange (METREX, EUROCITIES) and instruments to support the establishment of urban-rural partnerships (ZIT, RLKS), but the responsibility for setting up such partnerships lies with the local and regional actors. They need to identify the relevant thematic issues, learn about the principles of cooperation, and establish the organizational and decision-making structures to bring the partnership to life. Urban-rural partnerships cannot replace national spatial planning regulations, but they can complement each other and coordinate specific sectorial policies (Jacuniak-Suda et al., 2014;CEC, 1999;Kawka, 2013;Halamska and Stanny, 2021;Artmann et al., 2012). In the case of G<unk>dno, there is an existing local partnership. Both residents and rural and regional governments cooperate with each other. At the initiative of the village council and the Association for the Development of the Village of G<unk>dno, cooperation with non-regional cultural centers is developing. However, what is missing is the awareness of the need for cooperation and inclusion in the EU structures of smart villages. Below is a list of the essential benefits of implementing the smart villages concept that G<unk>dno may realize as a result of cooperation in the structure of EU smart villages. Each point is accompanied by "+" (yes) or "-" (no) in brackets, indicating whether G<unk>dno currently utilizes a similar approach. This essential benefits of implementing the smart villages concept are as follows: <unk> increasing the integration of the local community (+); <unk> encouraging young people to undertake project activities at the local level (+/-); <unk> increasing the involvement of older people in local activities (+); <unk> development of public-private partnerships (+/-); <unk> increasing human, material, and financial resources (+/-); <unk> developing social capital (+/-); <unk> reducing risks through diversification of project activities (-); <unk> improvement of the effectiveness of implemented projects, e.g., due to cost reduction (-/+); <unk> strengthening market position/competitiveness of local rural area (-); <unk> diversification of stakeholders of implemented projects (-); <unk> development of scientific and research cooperation (-); <unk> acceleration of socio-economic development e.g., through tourism (+); <unk> acceleration of knowledge and technology transfer (-).

The purpose of this text is to present an insignificant part of the field research (2021)(2022), about the reality of a small border village (Gądno), located in the West Pomeranian Voivodship, which (without knowing it) has partly and unconsciously joined the smart village path, implementing only some assumptions of the smart village concept and leaving others without attention. Design/Methodology/Approach: The manuscript presents a small part of field study activity of the Szczecin University stuff and students; research is based on an explanatory case study method, questionaries, and the desk-research method. The text is organized as follows: in the first part, a general picture of the actual state of agricultural land is outlined; then the literature review underlining the concept of partnerships and modern village organization is presented. In the next section, the methodology stressing the case study method has been placed. Then, there is a description of the village Gądno as a research results section followed by the discussion section. At the end of the work, the conclusion and practical implications are provided.The results of the case study contribute in the cognitive sense to the role of urban-rural smart village concept. There is proposed Gądno-village profile that combines information and communication technologies, social infrastructure and local and regional institutions which are dynamizing its economic, social, environmental, and cultural development. As well the role of trust between residents and the person who is the leader and village head is emphasized. Also there is mentioned, that Gądno-village is missing the need for cooperation and inclusion in the EU structures of smart villages. Practical Implications: List of the essential benefits of implementing the smart villages concept that Gądno may realize from the urban-rural cooperation, which can provide this village in the structure of EU smart villages. Each point is accompanied by the indicator whether Gądno currently utilizes or not a similar approach. Originality/Value: The results contribute in the cognitive sense to the smart village concept; shows an example of small village which (without knowing it) has partly and unconsciously joined the smart village path. They also show the importance of socioeconomic and spatial research in discovering the hidden, socio-cultural advantages of a village, that is not depopulated and has no problem with unemployment. The obtained results may be of interest to representatives of the public sector, as well as private entities seeking cooperation with local and regional authorities in the realm of smart village development.

INTRODUCTION By February 2023, more than 750 million SARS-CoV2 infections and 6.8 million COVID-19 associated deaths had been reported globally. 1 The development of vaccines against SARS-CoV2, which primarily protect against severe disease, has however greatly reduced both COVID-19related mortality and morbidity. 2 3 Global vaccination programmes were fast-tracked under the COVID-19 Vaccines Global Access (COVAX) initiative at the beginning of 2021. 4 Vaccination rates have, however, failed to meet the targets of vaccinating 70% of the world's population against COVID-19 WHAT IS ALREADY KNOWN ON THIS TOPIC <unk> Many countries have faced challenges when rolling out COVID-19 vaccines. Infrastructure, logistics, misinformation and vaccine hesitancy have been barriers to vaccine access and uptake globally. Vaccine nationalism by high-income countries has particularly affected countries in Africa and Asia, resulting in inequity between countries and regions. --- WHAT THIS STUDY ADDS <unk> Vaccine uptake among youth in Zimbabwe was more than 50% across all age-groups. Men, those with more education and those living under less socially deprived socioeconomic conditions were more likely to be vaccinated. Fear of side effects and myths circulating on social media were identified as barriers. Religion was less of a barrier than other studies reported, likely due to religious institutions' collaborations in COVID-19 vaccination efforts. --- HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY <unk> Vaccination campaigns should actively address specific concerns of communities, especially concerns around fertility and early death, and provide vaccines in easy-access and convenient locations. Involving community leaders in both education and vaccination efforts is pivotal given their trust and influence. on May 7, 2024 by guest. Protected by copyright. BMJ Global Health by mid-2022 set by the WHO, especially in the African continent. 5 By 3 February 2023, Africa has reported a vaccination coverage (receiving at least 1 dose of a COVID-19 vaccine) of eligible population of 46% and of total population of 27%, with coverage varying greatly from<unk>6% in Madagascar, Cameroon and the Democratic Republic of Congo to more than two-thirds of the eligible population in Rwanda and Liberia. 6 7 Differences are partly explained by availability of vaccines, infrastructural constraints and COVID-19 vaccine nationalism in high-income countries (HIC) leading to insufficient donations of vaccines to low-income and middle-income countries (LMICs). 8 In addition, vaccine hesitancy has been reported to be one of the main barriers towards meeting global vaccine coverage targets. [9][10][11] Studies conducted both in HICs and LMICs have reported widespread hesitancy towards receiving the COVID-19 vaccine, despite most of the population having previously received vaccines for other viruses. 12 13 Zimbabwe recorded its first case of COVID-19 in March 2020 and was subsequently among the first African countries to implement vaccination, with the first vaccine administered on 18 February 2021. 14 In contrast to other countries in the region, Zimbabwe procured COVID-19 vaccines (SinoPharm and SinoVac) through a bilateral agreement with China. 15 16 While non-availability of vaccines was a major barrier for vaccine uptake in other African countries, this was not the case in Zimbabwe. 16 Initially, healthcare workers and people working at borders were targeted for COVID-19 vaccination, followed by those with chronic conditions and essential workers such as teachers. 17 With the availability of more vaccine doses, eligibility was rapidly extended to the rest of the adult population (people aged 18 and over). In November 2021, vaccines were also made available to adolescents aged 16-18 years and in March 2022 to children aged 12 years and older. [16][17][18] The initial vaccine schedule was two doses of either SinoPharm or SinoVac taken 3 weeks apart. A third booster dose was introduced in January 2022. [19][20][21] The Zimbabwe vaccination campaign was administered through vaccination centres established in hospitals, clinics and outreach services. In addition, the vaccination campaign included educational programmes in schools, national mobilisation of frontline workers to assist in getting to hard-to-reach populations and nationwide health education broadcasting. 18 22 23 Once vaccination eligibility was extended to adolescents, the government partnered with organisations such as United Nations Children's Fund (UNICEF) to generate vaccination messaging targeting children and adolescents. 18 24-26 As of 3 February 2023, the Zimbabwean Government estimates that 44% of the total population have received at least one COVID-19 vaccine dose. With approximately 60% of the population being eligible (12+years), coverage among those eligible was 77%. 1 2 Despite the tremendous efforts put into the vaccination campaign, a considerable proportion of the general population remains unvaccinated and more recently vaccination coverage has stagnated. 14 While a Partnership for Evidence-Based Response to COVID-19 survey in September 2021 demonstrated that 82% of respondents in Zimbabwe were satisfied with the government's response, there was still some vaccine hesistancy. 27 Furthermore, there is a lack of evidence in the literature regarding uptake of COVID-19 vaccination among young people and potential causes of vaccine hesitancy in Zimbabwe. This study investigates self-reported COVID-19 vaccine uptake among 17 862 young people aged 18-24 years across three provinces (Harare, Bulawayo and Mashonaland East) in Zimbabwe and explores sociodemographic and -economic factors associated with uptake and reasons for possible vaccine hesitancy. --- METHODS Enrolment This study used data from a population-based survey which was conducted to ascertain the outcome of the CHIEDZA trial. CHIEDZA is a cluster randomised controlled trial conducted in three provinces (Harare, Bulawayo and Mashonaland East) in Zimbabwe investigating the impact of providing community-based integrated HIV and sexual and reproductive health services to young people aged 16-24 years on population-level HIV outcomes (NCT03719521). 28 Taking advantage of a large population-based survey being undertaken as the COVID-19 vaccination campaign was being rolled out, we sought to understand COVID-19 vaccination coverage among youth to inform future vaccination strategies. Youth were randomly selected from both the CHIEDZA intervention and control clusters (eight clusters per province, four being the intervention and four being the standard of care) using geographic information system methods. Each cluster was mapped and divided into street sections of approximately equal length. A random sample of street sections was selected, and all residents of those sections were enumerated. All residents aged 18-24 were assessed for eligibility to participate in the survey. The survey was conducted in Harare (October-December 2021), Bulawayo (January-March 2022) and Mashonaland East (April-June 2022) aiming to recruit 16 800 18-24 year-olds (5600 per province). The sample size was based on having sufficient power to ascertain the primary outcome of the trial, which was prevalence of viral suppression among young people living with HIV. A sample size of 700 per cluster was sufficient to detect a difference in the primary outcome of 21% at 80% power, assuming an HIV prevalence of 3%. 28 COVID-19 vaccines became eligible for youth (aged 18 and above) in March/ April 2021 and for 16-18 year olds in November 2021. 17 Sociodemographic data, self-reported COVID-19 vaccination and reasons for not being vaccinated were collected --- Data collection and management Participants viewed an information video about the study on a tablet screen prior to providing electronic consent. Survey data were collected onto electronic tablets using SurveyCTO (Cambridge, USA) and uploaded to the Biomedical Research and Training Institute (BRTI) server at the end of each day. Data were managed using Microsoft Access and stored in a password-controlled database, with access limited to define study personnel. Participants were identified only by ID number and no identifying data were maintained on the database. This study was performed in accordance with the study protocol, the Declaration of Helsinki as well as national and other regulatory guidelines. --- Outcome The outcome of this analysis was self-report of having received at least one dose of a COVID-19 vaccine. This was ascertained in the population-based survey with questions about COVID-19 vaccination: 'Have you had any dose of the COVID-19 vaccine?' with answer options being 'Yes-one dose', 'Yes-two doses', 'I have not had any doses'. At the time of the survey, everybody eligible for the survey (18-24 year olds) was also eligible for COVID-19 vaccination. A binary 'yes/no' response to vaccination was used for the analyses investigating associations. --- Risk factors The sociodemographic characteristics included in this analysis to understand factors associated with self-report of having received at least one dose of a COVID-19 vaccine were: age, sex, educational attainment, socioeconomic status defined through a principal component analysis, marital status and employment. Age was categorised, sex was binary (male/female) and education attainment was split into (1) none/primary, (2) form 4, (3) form 6 and (4) postsecondary. Marital status was grouped into (1) single, (2) married/living as if married and (3) divorced widows. Finally, employment was grouped into (1) none, (2) student, (3) employed (formal) and ( 4) employed (informal). --- Data analysis Data analysis was conducted using Stata V.16.1 (Stata, USA). A descriptive analysis was performed using proportions for categorical data, medians and means for continuous data, followed by a univariable logistic regression analysis to investigate the association between sociodemographic characteristics (sex, age, educational attainment, employment, marital status, socioeconomic BMJ Global Health status) and self-reported COVID-19 vaccine uptake. The outcome was self-report as having received at least one dose of a COVID-19 vaccine. Factors associated with the outcome in the univariable model were built into a multivariable logistic regression model, using only respondents with no missing data as only 22/17 682 (0.12%) were missing observations in the key covariates. To control for time, month of data collection was included in the model a priori. A likelihood ratio test was conducted for each variable to assess strength of evidence of association and reported as a p value. A sensitivity analysis was conducted including only participants from the control clusters of the CHIEDZA trial to understand whether the results of the analysis differed when including control clusters only as compared with using all data from intervention and control clusters. For the sensitivity analysis, a second multivariable logistic regression model was run using only participants from the control clusters. The socioeconomic status variable was created using a principal component analysis of ownership of assets (refrigerator, bicycle, car, TV, radio, microwave, cell phone and computer) and was then reported in quintiles (online supplemental table 4). 29 Analysis was performed for all provinces and then performed separately for each province as enrolment was conducted sequentially. Reasons for not taking up the vaccine were categorised according to the '5C's' as described by Razai et al: (1) confidence (safety and efficacy of the vaccine), ( 2) complacency (perception of low risk and disease severity), (3) convenience (barriers and access), ( 4) communication (sources of information) and ( 5) context (sociodemographic characteristics). 30 --- Patient and public involvement The information video about the study was codesigned with and piloted among youth. The study questionnaire was also piloted with youth. The study had a Youth Advisory Board that provided guidance on study design and conduct. An extensive public engagement programme was undertaken alongside the study that included a national crowdsourcing competition to (a) elicit young people's perceptions about health issues in their communities and (b) train youth as researchers through a mentored programme termed Youth Researchers Academy. 31 --- RESULTS Of the 18 682 randomly sampled eligible youth in the study communities, 17 682 (94.6%) provided consent to participate. Of these, 5849 (33.1%) were recruited in Harare, 5969 (33.8%) in Bulawayo and 5864 (33.2%) in Mashonaland East. The median age of survey participants was 20 years (IQR: 19-22), and 10 742 (60.8%) were women. This higher proportion of women reflected the community composition. Overall, 10 652 (60.3%) participants self-reported having received at least one dose of a COVID-19 vaccine, with 8316 (78.1%) having received two doses (figure 1). The proportion vaccinated was higher among males (n=4779, 68.9%) compared with females (n=5872, 54.7%) and increased with age (<unk>20: 57.5%, 20-22: 61.5%, >22: 62.2%) (table 1). Overall, Harare province, where data collection preceded the other provinces by 3 (Bulawayo) and 6 (Mashonaland East) months, had the lowest proportion vaccinated (42.6%) compared with Bulawayo (69.6%) and Mashonaland East (68.3%) (figure 1). The proportion vaccinated and double-vaccinated increased with each subsequent month during the survey in Harare and Bulawayo, but not in Mashonaland East. The major reasons given for not being vaccinated were lack of time, belief that the vaccine was unsafe and anxiety about side effects (figure 2). Men reported a lack of time as the main reason for not getting vaccinated. Women reported concerns related to side effects and safety as reasons for not getting vaccinated more frequently compared with men, especially infertility (10.7%). A total of 755 (4.3%) women in the survey were pregnant at the time of recruitment. Of these women, 297 (39.3%) reported having Univariable analysis (online supplemental table 2) showed an association between vaccine uptake and age, sex, educational attainment and socioeconomic status. The association between vaccine uptake and predictors remained in the multivariable analysis (table 2). Men (OR 1.69, 95% CI 1.59 to 1.80), older youth (20-22: OR 1.06, 95% CI 0.98 to 1.15, >22: OR 1.12, 95% CI 1.04 to 1.21), those with higher educational attainment (form 4: OR 1.79, 95% CI 1.39 to 2.30, form 6: OR 3.56, 95% CI 2.72 to 4.66, postsecondary: OR 4.34, 95% CI 3.27 to 5.76) and higher socioeconomic status (second quintile: OR 1.06, 95% CI 0.96 to 1.17, third quintile: OR 1.12, 95% CI 1.01 to 1.23, fourth quintile: OR 1.17, 95% CI 1.06 to 1.29, least poor 20%: OR 1.32, 95% CI 1.20 to 1.47) were more likely to be vaccinated. Overall results and results stratified by province were comparable. The sensitivity analysis conducted with the same exposure variables using only the control clusters of the CHIEDZA intervention yielded similar results (online supplemental table 3). --- DISCUSSION This study found a COVID-19 vaccine coverage among young people aged 18-24 years of 69.6% in Bulawayo, 68.3% in Mashonaland East and 42.6% in Harare. Despite not being a high-risk group, the national vaccination campaign reached them effectively. Vaccine uptake was however inequitable. Those who were male, older, more educated and of higher socioeconomic status were more likely to report COVID-19 vaccination, which is in line with studies conducted in HICs, though the specific reasons may be different. Education attainment and male sex were the strongest predictors. The difference in COVID-19 vaccine coverage among youth across the provinces is partly explained by the staggered timing of the survey. The national vaccination campaign started in February 2021 prioritising front-line workers. Vaccination eligibility was extended to all adults in March/April 2021 and 16-18 year-olds became eligible in November 2021 (1 month after study recruitment started in Harare). At the time the survey was completed --- BMJ Global Health in each province, the provincial vaccine coverage in Zimbabwe was 25.3% (Harare), 31.7% (Bulawayo) and 38.6% (Mashonaland East). 32 Importantly, in the context of COVID-19 vaccination, coverage has been defined as the percentage of the total population that is vaccinated and includes children even though they may not be eligible. 33 At the end of the survey, vaccine coverage using the total eligible population (16+ years) as denominator was 73.8% nationally and thus comparable to coverage among youth in Mashonaland East. 2 In the full adjusted model, men had 1.69 times the odds of having received the vaccine compared with women. Other studies have found higher proportions of men compared with women reporting intention to getting the COVID-19 vaccine. [34][35][36] These studies have also highlighted that the difference in intention to getting vaccinated against COVID-19 between men and women is less about increased health-seeking behaviour in men and more about decreased health-seeking behaviour among women, which is consistent with existing literature on general vaccine hesitancy. 37 38 This difference is likely due to specific gender differences in risk aversion and potential side effects. Reasons for not getting vaccinated in our study were mainly related to confidence, that is, regarding safety and efficacy of the vaccine, especially among women. In this survey, 20.1% of women who were not vaccinated said they were afraid of side effects in general and 10.7% said they were afraid of infertility specifically. Furthermore, 66/458 (14.4%) of unvaccinated pregnant women reported fertility-related fears as a barrier, despite the WHO having recommended the use of SinoPharm in pregnant women. 34-36 39 This disparity may be due to confusing communication regarding pregnancy and breast feeding at the start of the vaccination campaign, including information by official sources and on social media. 40 41 The infodemic about vaccines spread through social media has undoubtably played an important role in Zimbabwe given the high proportion of individuals (both men and women) who reported being afraid of dying within 2 years of vaccination. This myth relates to a widely and globally circulated text message meme claiming that French virologist Luc Montagnier had said all vaccinated people will 'die within 2 years'. 42 Fear of side effects was less frequent among unvaccinated men compared with women, and a higher proportion of men did not feel at risk (felt 'young and healthy') or said they were too busy for vaccination. This is despite a vaccination campaign that tried to bring vaccines to the population by offering transport incentives such as transport money, decentralising vaccination to polyclinics and providing vaccine outreach services. 16 Vaccination rates varied by socioeconomic status and particularly educational attainment, pointing towards health inequity. Other studies have reported that poorer and less educated people experience more barriers to vaccination. 43 44 Those with lower educational attainment --- BMJ Global Health are less likely to have access to accurate information and the vaccine information itself might be inaccessible in terms of language, content and format, especially if this is provided in a written format. 45 Public health information may also include jargon which discriminates against those of lower education level. Those of lower socioeconomic status were also more likely to experience adverse events such as a cut in household income, inability to access food, higher disease risk or loss of work than those in higher quintiles, which affects people's ability to receive a vaccine. 46 In this study, only a small proportion of individuals said they were not vaccinated because of their religious beliefs. This is in contrast to the results from studies conducted in other countries in Africa where religious beliefs were among the most frequent reasons given for vaccine hesitancy. [10][11][12] In Zimbabwe, the Apostolic church, a Pentecostal Christian denomination, has an estimated 3.5 million followers mainly among poorer and rural households. Children of members of the Apostolic church have been found to have low childhood vaccination coverage. 47 Religious affiliation was not collected in the survey, but membership of the Apostolic church is less common in urban and periurban settings, which may explain some of our findings. Also, the Ministry of Health and Child Care, supported by UNICEF, actively reached out to churches including the Apostolic Church to support vaccine education and public engagement. 48 The Apostolic Women's Empowerment Trust started a COVID-19 awareness programmes in 2021 with the aim to address vaccine hesitancy among those of Apostolic faith. 49 Strengths of this study include a large representative sample of young people from three provinces at the time of the COVID-19 vaccine rollout. Participation in the survey was high and results were comparable across all three provinces. The main limitation of this study is that COVID-19 vaccination status was self-reported and social desirability bias may have resulted in overestimation of vaccination coverage. However, the survey included a range of potentially sensitive questions, on topics including sexual and reproductive health, and substantial efforts were invested in training the survey team. The survey was conducted in urban and periurban settings only and vaccination coverage may be very different among young people living in rural areas. There were no questions specifying the type of vaccine received or the date of vaccination. Another limitation was that the timing of the survey relative to the COVID-19 vaccination campaign differed between the three provinces. Furthermore, the multivariable logistic regression used univariable analysis screening methods to select the covariates for the model. This method may have erroneously excluded covariates in the full model and have led to biased estimates. Finally, while we asked young people to provide reasons for not being vaccinated, more detailed questions on beliefs, myths and sources of information might have provided better understanding. Further qualitative research investigating the barriers towards vaccination, especially among women, would strongly contribute to the findings reported in this study. --- CONCLUSION This study showed inequitable COVID-19 vaccine coverage among young people aged 18-24 years in three provinces in Zimbabwe (Harare, Bulawayo and Mashonaland East). Vaccination rates were lower among women, people with lower educational attainment and among those living in more deprived socioeconomic circumstances. Fear around death and infertility were among the main reasons for vaccine hesitancy. Therefore, we recommend that national vaccination campaigns should include a major focus on health education, especially for women and those with less education, and particularly around infertility and death. Information that is clear and consistent provided by trusted sources is crucial, and in this way, social media can be used in a positive manner to combat misinformation. Furthermore, campaigns should focus on community and religious leaders as they have a strong impact on their respective groups. No single programme is likely to address vaccine hesitancy in this population and thus strategies should address these inequities by understanding concerns and tailoring vaccine campaigns to specific groups which may be at a socioeconomic disadvantage instead of a one-sizefits-all approach. --- Competing interests None declared. Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details. Patient consent for publication Not applicable. --- Ethics approval The study received ethical approval from the Medical Research Council of Zimbabwe (MRCZ) (MRCZ/A/2387), the BRTI Institutional Review Board (IRB) (AP149/2018) and the London School of Hygiene and Tropical Medicine (LSHTM) ethics committee (16124). Provenance and peer review Not commissioned; externally peer reviewed. Data availability statement Data are available on reasonable request. The data will be made available on LSHTM Data Compass within 12 months of publication. Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Introduction COVID-19 vaccine acceptance research has mostly originated from high-income countries and reasons why youth may not get vaccinated may differ in low-income settings. Understanding vaccination coverage across different population groups and the sociocultural influences in healthcare delivery is important to inform targeted vaccination campaigns. Methods A population-based survey was conducted in 24 communities across three provinces (Harare, Bulawayo and Mashonaland East) in Zimbabwe between October 2021 and June 2022. Youth aged 18-24 years were randomly selected using multistage sampling. Sociodemographic characteristics, COVID-19 vaccination uptake and reasons for non-uptake were collected, and odds of vaccination was investigated using logistic regression. Results 17 682 youth were recruited in the survey (n=10 742, 60.8% female). The median age of participants was 20 (IQR: 19-22) years. Almost two thirds (n=10 652, 60.2%) reported receiving at least one dose of COVID-19 vaccine. A higher proportion of men than women had been vaccinated (68.9% vs 54.7%), and vaccination prevalence increased with age (<19 years: 57.5%, 20-22: 61.5%, >23: 62.2%). Lack of time to get vaccinated, belief that the vaccine was unsafe and anxiety about side effects (particularly infertility) were the main reasons for not getting vaccinated. Factors associated with vaccination were male sex (OR=1.69, 95% CI 1.58 to 1.80), increasing age (>22 years: OR=1.12, 95% CI 1.04 to 1.21), education level (postsecondary: OR=4.34, 95% CI 3.27 to 5.76) and socioeconomic status (least poor: OR=1.32, 95% CI 1.20 to 1.47). Conclusion This study found vaccine inequity across age, sex, educational attainment and socioeconomic status among youth. Strategies should address these inequities by understanding concerns and tailoring vaccine campaigns to specific groups.

Introduction --- C urrent migration dynamics have contributed to the growing diversity of populations in Europe and worldwide. 1 During 2015, 2.4 million people from non-EU countries (including many asylum seekers) migrated to one of the EU-28 member states. 2 This situation has become a key public health issue as it challenges healthcare systems to meet the needs of increasingly heterogeneous populations. It has become critically important for health policymakers, planners and equality bodies-across all nations-to obtain and use high-quality data that adequately captures the composition of populations to inform measures aimed at identifying and reducing disparities. 3,4 In 2015, the Robert Wood Johnson Foundation (RWJF) (www. rwjf.org/), in the USA, instigated a multi-component project to examine ethnic/racial data collection in the USA. The US recognize six categories of ethnicity for data collection, and there are a large number of categories not officially reported. 5,6 These categories do not provide detailed enough information to identify variations in health needs, and the use of these categories can mask existing inequalities in health, healthcare and social determinants. The RWJF decided to explore how these data could be disaggregated into more 'granular' categories, meaning more detailed/finer categorization. For example, the 'Asian' category used in the US encapsulates a large, heterogeneous population which could be disaggregated to origins from east of Istanbul through to Japan. 7 To supplement and inform this work, the RWJF commissioned our Heterogeneity/Granularity in Ethnicity Classifications (HGEC) project to gather perspectives, from within Europe and internationally, on approaches to collecting granular ethnicity data. This paper presents the HGEC European overview. It is the first study to examine the granularity of ethnic classification in the European Union, though a prior study has looked into the different methodologies for collecting ethnic data in Europe. 8 Previous research has examined the availability of official international data sources which collect data on ethnicity. 9,10 A global survey of ethnic enumeration in the 2000 national censuses found that 63% of countries included some ethnic categorization. 9 In Europe, projects have also examined official data sources for migrant and ethnicity/race health information (e.g. the Migrant Integration Policy Index 'MIPEX' and Migrant and Ethnic Health Observatory 'MEHO') 11,12 and these demonstrated a very mixed landscape. In Britain, e.g. there is support for the collection of ethnicity data and the health of minority ethnic groups is a government priority within population health policy. 13 The British census office noted that ethnicity data was collected because of a recognized need for reliable information about diverse socioeconomic factors, and these data have been used to study inequalities in health. 14 There is also supportive legislation which highlights the crucial role of data in supporting anti-discrimination policies. 15,16 Conversely, much debate surrounds the benefits and risks of the official collection and use of ethnicity data within other parts of Europe, as it is regarded as sensitive information. 4,15 In France, e.g. it is feared data could be misused to create and maintain divisions between majority and minority groups. By law, France does not collect ethnicity in their census. However, under the European Data Protection Directive, no prohibition of ethnic data collection exists and France's position is their interpretation of the law. 4 Resulting from these tensions, collection of ethnicity data either remains inadequate in most of Europe because countries do not systematically collect data, or because what they do collect is inadequate for health research, policy and practical purposes. 8,17,18 Even in countries where relatively granular data are collected, there remain limitations in capturing the diversity required to inform analyses of health inequalities. 19 --- Methods We used official data sources from the EU-28, 20,21 to give an overview of the collection of granular ethnicity at the European level, using only official EU member states. Supra-national data sources were also consulted to provide greater insight to our findings, but were not used as a source of primary data. For the purpose of this study, we considered the four UK countries individually: England, Scotland, Wales and Northern Ireland. --- Research questions (i) From official sources within EU-28 countries, which countries collect data on ethnicity, and what degree of granularity is there within their classifications? (ii) For those countries that do collect ethnicity data, what approaches to terminology are used within their categorizations? --- Identification of literature and data extraction We identified literature by searching the sources listed in box 1, devised in consultation with two expert librarians. Our Research Fellow (NV) screened titles and abstracts found in electronic databases, documents from citations and key websites according to agreed inclusion and exclusion criteria (box 1). As the HGEC findings are to inform the wider work of the RWJF, we organized our results according to the standards for the classification of federal data on race in the US census bureau, as well as in the US Office of Management and Budget, which have a minimum of six categories. 5,6 Granularity was determined by the number of ethnic categories a country collects. We considered that countries with one to six 'tick box' categories had some granular approach; while countries with more than six 'tick box' categories and countries with a write-in option (allowing any free text answer) were considered as having a more granular classification. --- Expert peer review An international expert group was organized to provide feedback on preliminary results. The results of this component of the project were internally peer reviewed by two members of this expert group. --- Results --- Collection of ethnicity data and degree of granularity We identified six countries whose official census had one to six ethnic categories (table 1); the maximum number of categories being six (Estonia). Most countries included a write-in option. Two countries (Bulgaria and Slovenia) included an option where respondents could choose not to answer. For countries such as Estonia, Lithuania and Slovenia, the census question on ethnicity was based on nationality, whereas Cyprus and Bulgaria included ethno-religious groups (e.g. Maronite) and Roma people, respectively. We found eight EU-28 countries with more than six categories (table 2), the maximum being 19 (Scotland and Hungary). Four of these countries were within the UK: Scotland, England, Northern Ireland and Wales. These four countries offered a wide range of options for disaggregating the White population group. In addition, the Scottish, English and Welsh census had relatively disaggregated classification for Asian and African, Caribbean or Black groups. In contrast, the Republic of Ireland and Northern Ireland provided limited options for these groups. There is a rapidly increasing population of people globally who identify as mixed-ethnicity and this has been included as a category in Scotland, England, Northern Ireland and Wales. Scotland's mixed category is a write-in option, compared with England and Wales which included disaggregated 'tick boxes' within the Mixed/ multiple ethnic group category (e.g. White and Black Caribbean) and also a write-in option. All UK countries provided a Gypsy/Traveller category ('Irish Traveller' in Northern Ireland) and the Republic of Ireland had 'White Irish Traveller'. Hungary, Poland and Slovakia included an option for the Roma population. However, these categories may still consist of people with very diverse ethnic backgrounds, e.g. the Roma population sub-groups 22 and Hungarian sub-groups (e.g. Romungro, Bea <unk>s and Lova <unk>ri). Regarding the rest of the countries, we found three countries (Latvia, Romania and Czech Republic) provided solely a 'write-in' option to self-identify using free text. In the 2011 Romanian census, the ethnicity question asked 'What ethnic group does the person consider that he/she belongs to?,' stating that 'Each person is free to express his or her opinion, without any constraints.' 21 No English abstract was available Data were collected before a respective state was a member of the EU The country was an overseas territory of an EU member state or not part of the 28 member countries of the EU The remaining 14 countries (table 3) did not record ethnicity and instead recorded related variables, such as country of birth, parents' country of birth, citizenship or nationality. --- Approaches to terminology used within ethnic categorizations In countries where ethnicity was collected, operationalization of the concept varied, as reflected in both the questions asked and the response categories. Generally, the question asked for 'ethnicity' combined with other terms, such as 'ethnic affiliation' (Poland), 'ethnic nationality' (Hungary), 'ethno-religious' group (The Republic of Cyprus) and 'ethnic or cultural Background' (Republic of Ireland). Response categories also used combinations of concepts. The four UK countries based their classification on a concept of ethnicity with elements of ancestry, geographical origin and skin color. Most other countries based their categories on national identities 17 (e.g. Hungary), both within their population and the main immigrant groups (Bulgaria, Estonia, Lithuania and Croatia). Another approach was the use of ethno-religious groups, as seen in Poland, Slovakia and the Republic of Cyprus (e.g. Maronite and Latin). Some countries combined these approaches to attempt to measure how people perceive themselves. In Poland, e.g. the categories included a combination of nationality (Polish, Belarussian, Czech, Lithuanian, German, Armenian, Russian, Slovakian and Ukrainian), ethnicity (Karaitic, Lemko), language/religion (Tatar, Jewish) and ethnic minority groups (Romany). While, uniquely, the 2011 census of Slovenia gave the option of non-disclosure, with the choices of 'I'm nationally/ethnically indeterminate,' and 'I don't wish to answer this question.' 21 --- Discussion --- Summary of main findings This study summarizes current approaches to the collection of granular ethnicity data within EU-28 countries. Our results showed: (i) Although there is reasonable attention to the diversity of ethnic groups in data collection, a granular approach does not predominate within EU-28 classifications. (ii) Where ethnicity is collected, it is conceptualized in different ways and diverse terminology used. (iii) A write-in option provides the most granular approach. (iv) Almost half of the countries did not collect data on ethnicity, but did collect related variables that could be used as a proxy. --- Generalizability This project was restricted to the EU-28. Although this work is specific to these countries' contexts, the data collection and approaches to ethnic classification are likely to be generalizable to other settings. --- Strengths and limitations Strengths of the HGEC project include the breadth of overview and the attention given to exploring the granularity of ethnic classifications across the EU-28. In this paper, we have also contextualized our findings in relation to the international literature. Examination of additional information sources could have augmented our findings, particularly supra-national data sets and country-specific health data sets (surveys and routinely collected data). This may have revealed greater complexities, e.g. situations where it is forbidden to record ethnicity on healthcare records, explicitly to prevent discrimination, and investigation of how researchers approach such situations. However, it was too large a task to comprehensively access these data for the number of countries covered in this overview. It would have also been interesting to explore intersectionality (e.g. of ethnicity, gender, socioeconomic status) and multiple discrimination, but this was also beyond the scope of this project. The HGEC project is part of a larger international project for the RWJF in which we also undertook more in-depth reports of data sources for seven countries (three within Europe), a summary of which will be published by PolicyLink (http://www.policylink.org/ our-work/community/health-equity/data-disaggregation). --- Relationship to existing knowledge and practice --- Ethnicity data collection Research has demonstrated inequalities in health across diverse ethnic groups. 7,[23][24][25] Despite the recognized need to identify ethnic groups to promote equity in health policy and planning purposes, we found that granular classifications do not predominate within the EU-28. It appears more commonplace, and perhaps more accepted, to collect variables that could be used as a proxy, such as nationality, country of birth and parents' country of birth. 26 Some studies have used these variables to identify the generation of migration (e.g. second generation, children of migrants born in the receiving country) which, in relation to health, can influence health behaviors and risks of disease. However, for successive generations this approach fails to adequately capture population diversity. Routinely collecting ethnicity data within healthcare settings is a preferred model for analyzing health/healthcare inequalities. 27 However, even in countries which have collected ethnicity information for some time in their census, the routine collection of these data is still relatively new. Their analyses have primarily relied on research using data linkage of census and healthcare datasets. 28,29 At this time, the census classifications likely best reflect each country's approach to ethnicity data collection. --- Conceptualizing ethnicity and diverse terminology We found that censuses and population registers which did measure ethnicity, do so in very diverse ways. Even when examining countries with comparable numbers of categories, there was no consistent approach to the classifications used, except within the UK, where public bodies, e.g. the NHS, are bound by law to the collection of equality data. Some of the most disaggregated categorizations, such as in the UK and Republic of Ireland (which does ask for 'ethnic or cultural background'), despite multiple response options, still lack a strong cultural dimension (seen in countries who incorporate language and religion) which could help reflect 'belonging to a group' and also current conceptual shifts from biological conceptions of ethnicity, to the cultural understanding of group identity. 30 Census classification faces even greater challenges due to the growth of mixed or multiple identities within population groups. [31][32][33] Previous research has shown that 'dual identification and multicultural environments may be associated with a positive sense of racial identity,' and also'mixed' population groups have been demonstrated in some studies to exhibit poorer health outcomes. 14,34,35 Some countries (e.g. England, Scotland and Wales) include a mixed group category. Scotland had a single write-in category compared with England and Wales, which had four categories (three exact combinations and one write-in option). However, further work is needed to develop approaches to this categorization, 7,33,36 as well as how to analyze data for'mixed' populations and unpack the meaning of findings for a group that is inherently so diverse and increasingly heterogeneous. A write-in option as the most granular approach Three countries (Latvia, Romania and the Czech Republic) were found to offer a 'write-in' option for ethnicity and were exclusively free text, with no predetermined categories to choose from. This approach offers the most scope for granularity. However, a sole write-in option can create difficulties for respondents in understanding the question, with no set categories to refer to, leading to poorer quality data and lower response rates. It may also be problematic for data interpretation due to the vast number of potential responses and the subsequent decisions required for statistical analyses. 37 These data are often only collated and analyzed at an aggregate level, which negates the value of this granular approach, and removes the element of self-identification. Thus, this method will only be valid if granularity is maintained when high-quality data are analyzed, reported and used to inform health care and policy. --- Countries collecting variables that could be used as a proxy Almost half of the EU-28 countries did not collect ethnicity, but did collect related variables. As mentioned earlier, in France there is a refusal to collect ethnic data due to the judicial interpretation of data collection which may be sensitive. Other countries may not have considered collecting ethnicity, and some view related variables as more informative. For example, the Netherlands prefer the objective measure 'country of birth' over what they view as a more subjective measure of ethnicity. 26 Collecting nationality is also recognized as important, particularly to understand healthcare access for migrant populations, 11 though the use of nationality as a variable can be ambiguous; sometimes referring to origins and other times to legal citizenship status. Potentially, collecting a cluster of variables (e.g. nationality, country of birth, language spoken and ethnicity) presents the most comprehensive approach for assessing health inequities in diverse population groups, selecting variables where appropriate for different research enquiries. --- Potential factors influencing the collection of ethnicity and categories/classifications Diverse approaches to the collection and classification of ethnicity data relate, in part, to governmental influence on ethnic enumeration and an individual country's ideology and perspectives on operationalizing these data. 10 Some country-specific contextual factors which may influence their perspectives include; having, or not having a long history of immigration diversity; low ethnic density; politics and concerns among the policy making community; historical events, including past conflict and persecution; interethnic relations and ongoing migration patterns. The position of some countries is that collection of ethnic data could lead to the essentialization of ethnic groups in forcing individuals to identify solely with one ethnicity, and to the abuse of data and racial discrimination. It is an interesting dichotomy that explanations both for the absence of ethnic statistical data in some EU countries and its promotion in others might be used in efforts to in the design legislation to prevent discrimination and promote cohesion. 8 There are also conceptual and situational challenges for official population registers/census. The conceptual issue is a lack of shared understanding of what determines ethnicity, or why such data is important for public health. The situational issue refers to potential stigma and fear associated with disclosing ethnicity, the degree of which may vary according to the context in which data are collected, meaning that figures from different sources may vary widely and official data may not be reliable, especially at a granular level for vulnerable groups (e.g. Roma population). 38 --- How this work can influence future research and practice Our findings concur with previous explorations 8,9 of ethnicity data collection within Europe, showing an extremely varied approach, ranging from some countries not collecting ethnicity, to others taking significant consideration of granularity. To achieve a situation where ethnic inequalities in health and health care can be identified and addressed through public health research and practice, there needs to be a more consistent and granular approach to the collection of ethnic data within Europe. Ethnic classifications and data use in Europe do not reflect the current dynamics of population diversity (e.g. fixed homogenous ethnic categories in the census) and this creates gaps in understanding the sociocultural and political factors that influence people's health and quality of life. Failure to include disaggregated data hinder efforts to understand, routinely monitor and address inequities in health in Europe. The use of more disaggregated ethnicity data, in conjunction with other related variables, would help the public health community to unveil the 'invisibility' of diverse population groups. However, it is pivotal to recognize that the diversity of approaches in ethnic classifications in the EU-28 reflect multiple countryspecific contextual factors. This complexity impedes the idea of advocating for a unified approach to the collection of ethnicity data for population health purposes within Europe, and as a result, the persisting diversity of approaches will continue to create obstacles for cross-border research and international comparisons. Nevertheless, it may be possible to promote agreement of some overarching principles and the idea of increasing the collection, and granularity, of ethnicity data in ways which are appropriate to each country's context. --- Conflicts of interest: None declared. --- Key points Ethnicity is conceptualized in different ways within European official data collection, using diverse terminology, and a granular (fine detailed) approach to ethnicity classification does not predominate within EU-28. Almost half of the EU-28 countries collect related variables that could be used as a proxy. The complexity of country-specific contexts within EU-28 impedes a unified approach to ethnicity classification, but it may be possible to agree on some overarching principles.

Background: Identifying ethnic inequalities in health requires data with sufficiently 'granular' (fine detailed) classifications of ethnicity to capture sub-group variation in healthcare use, risk factors and health behaviors. The Robert Wood Johnson Foundation (RWJF), in the USA, commissioned us to explore granular approaches to ethnicity data collection outside of the USA, commencing with the European Union. Methods: We examined official data sources (population censuses/registers) within the EU-28 to determine the granularity of their approach to ethnicity. When ethnic information was not available, related variables were sought (e.g. country of birth). Results: Within the EU-28, we found 55% of countries collected data on ethnicity. However, only 26% of these countries (England, Wales, Northern Ireland, Scotland, Republic of Ireland, Hungary, Poland and Slovakia) had a granular approach, with half of these being within the UK. Estonia, Lithuania, Croatia, Bulgaria, Republic of Cyprus and Slovenia collected one to six categories. A 'write-in' option only was found in Latvia, Romania and the Czech Republic. Forty-five percent of countries did not collect ethnicity data but collected other related variables. Conclusions: (i) Although there is reasonable attention to the diversity of ethnic groups in data collection, a granular approach does not predominate within EU-28 classifications. (ii) Where ethnicity is collected, it is conceptualized in different ways and diverse terminology is used. (iii) A write-in option provides the most granular approach. (iv) Almost half of the countries did not collect data on ethnicity, but did collect related variables that could be used as a proxy.

INTRODUCTION A community or society's social capital is made up of its networks, ties, and trust. The intangible asset in issue is crucial for promoting social cohesiveness, accelerating economic development, and improving people's general well-being both individually and collectively (Villalonga-Olives et al., 2022). Social capital may be divided into two different categories: interpersonal connections inside a certain group or community, known as "bonding social capital," and interpersonal connections between other groups or communities, known as "bridging social capital" (Quang Trinh et al., 2023). The cohabitation of rural and urban areas in Indonesia is notable for each displaying distinctive social dynamics (Langer et al., 2023). The idea of social capital is crucial for determining the social structure of these communities and influencing their development. The existence of established institutions and organizations within urban environments cultivates a sense of trust and collaboration among individuals and groups (Park, 2023). Consequently, this phenomenon results in heightened social mobility and expanded economic prospects. Conversely, rural regions in Indonesia exhibit a distinctive feature of tightly-knit communities that heavily depend on social capital formed through bonding (Jääskeläinen et al., 2023). Robust family and kinship networks often unite these communities, share cultural values, and adhere to traditional norms. Establishing a sense of belonging and trust within a community is of utmost importance for the efficient operation of rural regions, as it enables the facilitation of cooperation, the sharing of resources, and the provision of mutual support (Mittenzwei et al., 2023). In the rural areas of Indonesia, social capital can be witnessed through mutual assistance groups commonly referred to as "gotong royong" (Kushandajani & Alfirdaus, 2019). These collectives Prastyo, R., E., Wisadirana, D., Rozuli, A., I., Hakim, M., L. (2024). Social Capital's Impact on Indonesia's Urban and Rural Areas 3 convene voluntarily to engage in various activities, including constructing public works, agricultural endeavours, and coordinating cultural festivities (Hutagalung et al., 2022). Go to Gotong Royong fosters social cohesion and collective responsibility through its efforts to address shared needs and challenges the community encounters. However, it is important to remember that rural places can display strong amounts of social capital that fosters bonds. They typically struggle, though, with a lack of the bridging social capital that is common in metropolitan contexts (Puskarova, 2022). A greater sense of isolation may be caused by limited access to resources, information, and opportunities in rural areas (Maeder et al., 2023). This situation can make it difficult for people to interact with and gain benefits from outside networks and support systems (Henderson et al., 2020). As a result, rural areas could have trouble luring investors, improving their infrastructure, and gaining access to essential services. The creation of social capital that unites rural and urban communities is crucial for allowing the sharing of resources and knowledge and for supporting the in rural regions, social capital frequently derives from robust familial and communal affiliations. Collaborative efforts are commonly observed among individuals who engage in mutual assistance and cooperation, particularly in construction projects or agricultural cultivation (Mittenzwei et al., 2023). This type of social capital is necessary for rural communities to function effectively and care for one another (Antinyan et al., 2022). Through examining social capital in urban and rural contexts, valuable insights can be gained regarding the functioning and factors contributing to the prosperity of communities (Karnik & Peterson, 2023). This knowledge can inform decision-making processes and policy formulation, fostering community support and enhancing resilience (Beckham et al., 2023). --- THEORETICAL FRAMEWORK The social capital of a community or culture is comprised of its trust, relationships, and networks. The phenomenon being investigated relates to interpersonal relationships and friendships (Bashar, 2022). Two kinds of social capital can be distinguished: bonding social capital, which is focused on the bonds formed inside a particular group, and bridging social capital, which describes the connections formed between different groups (Del-Castillo-Feito et al., 2022). In Malaysia, social capital is crucial for community development, encompassing networks, relationships, and trust formed through informal encounters (Zainoddin et al., 2022). It fosters cooperation, collaboration, and collective action. It can be classified into bonding, bridging, and linking types, with bonding involving close connections, bridging involving connections from different social groups, and linking involving diverse authority or status (Claridge, 2018). In Indonesia, social capital is essential for the growth of communities in both urban and rural areas. In urban regions with diverse populations, social capital plays a crucial role in fostering intergroup ties and teamwork among individuals of different backgrounds (Marschlich & Ingenhoff, 2021). The study of social capital in urban and rural Indonesia is important because it helps understand the methods by which various groups cooperate and help one another (Bruno et al., 2021). We can make choices and implement policies that enhance and assist communities when we have a solid understanding of social capital. --- RESEARCH METHODS This article employs a literature review methodology involving understanding and studying theories from various sources. Zed ( 2004) outlines four stages in this process. Koopman (2023) suggests using empirical techniques to examine social capital in urban and rural Indonesia. Data collection involves searching and extracting information from diverse references, followed by critical analysis, including paraphrasing for clarity and support. The approach allows for a detailed exploration of social dynamics and interpersonal relationships (Daly et al., 2023). Qualitative methodologies can provide indepth insights into context-specific factors (Allen et al., 2022). The study aims to enhance the understanding of social capital in Indonesia's urban and rural areas (Perez-Ferragut et al., 2022), with potential implications for decision-making and policy development. --- RESULT AND DISCUSSION --- A COMPARISON OF SOCIAL CAPITAL LEVELS IN INDONESIAN URBAN AND RURAL AREAS The results of the study will provide a thorough analysis of social capital levels in Indonesia's urban and rural areas. It might include information on the size and makeup of social capital networks, personal links, and degrees of trust in distinct groups (Shin, 2022). The study examines how social capital differs and overlaps in urban and rural areas, as well as how social capital affects chances for employment and general well-being in each of these environments (Dasanayaka & Matsuda, 2022). The study's findings may also point to specific challenges and opportunities for boosting social capital in Indonesia's urban and rural areas. In both urban and rural locations, social capital is developed and maintained through a variety of mechanisms that are examined in this essay. The study looks at the numerous elements that have an impact on the growth and maintenance of social capital in Indonesia's urban and rural communities. It could examine how diverse socioeconomic elements, such as local associations, cultural norms, and social networks, affect the development and expansion of social capital (Colozza et al., 2023). The findings of this study may highlight the importance of a number of elements, such as social cohesiveness, reciprocity, trust, and shared values, in fostering social capital in both urban and rural contexts. The present discourse revolves around the challenges and opportunities of enhancing social capital. The study aims to evaluate the obstacles and potential advantages associated with enhancing social capital in urban and rural areas of Indonesia (Huang et al., 2023). It may encompass the identification of various obstacles to the formation and sustenance of social capital, such as disparities in wealth, exclusion from social networks, or disparities in cultural norms (Ieamkaew et al., 2020). Additionally, it involves recognizing potential avenues for enhancing social capital through initiatives rooted in local communities, interventions in policy-making, or programs aimed at strengthening individual and collective capabilities (Ver Steeg, 2022). The results of this study may offer valuable insights into the various strategies and approaches that can be employed to enhance social capital and foster community development in both urban and rural settings (Slijper et al., 2022). There is also social capital in an ethnic identity that is used to foster nationalism and tolerance, as shown in a comprehensive study of the The study compares social standing measures in Indonesia's urban and rural areas. (Waliuzzaman & Alam, 2022). It would be necessary to examine a variety of variables, such as the level of trust, the number of neighborhoods groups, the frequency of social encounters, and the strength of social networks in urban and rural settings. (Puppala et al., 2023). The findings of this study may provide insight into how social capital differs and is comparable in urban and rural settings. (Czy<unk>ewski et al., 2021). Additionally, it may offer valuable insights into the distinct challenges and opportunities associated with forming and maintaining social capital in these communities (Abdullah & Rahman, 2021). In other research findings, it was found that in the field of urban agriculture initiatives, social capital can be used to build networks and mobilize resources for sustainable development (Nazuri, 2023). This comparative research can offer helpful insights for creating strategies and actions that are specifically tailored to match the unique requirements and features of Indonesia's urban and rural areas. (Ebewore, 2020). This passage elucidates the theoretical frameworks, methodologies, and findings about examining social capital in urban and rural regions of Indonesia (Stegelmeier et al., 2020). Social capital is a concept that encompasses the various connections, networks, and relationships that individuals possess within a given community (Tóth et al., 2022). These social connections can be utilized to access valuable resources and receive assistance and support (Piras et al., 2021). The text discusses participant selection and sampling strategies in research. In quantitative studies, random sampling is used to ensure a representative sample (Kassie et al., 2022) The text gives a summary of many theoretical models important for understanding the idea of social capital. (Fielke et al., 2022). These frameworks include Putnam's theory, Coleman's theory, Bourdieu's theory, and Lin's theory. Each framework offers distinct viewpoints on social capital, highlighting elements such as trust, social networks, social class, and social structures (Richmond & Casali, 2022). The passage additionally addresses various research methodologies that can be employed to examine social capital in Indonesia, encompassing quantitative, qualitative, and mixed-methods approaches (H. Y. Li et al., 2019). The collecting and statistical analysis of numerical data are necessary for quantitative research. Comparatively, qualitative research entails gathering nonnumerical data to fully understand social capital. (Lou et al., 2023). Mixed-methods research combines quantitative and qualitative procedures (Dickey et al., 2022). The data analysis techniques mentioned include statistical analysis for quantitative data and thematic analysis, content analysis, and grounded theory methods for qualitative data (Bellavista et al., 2022). The study also emphasizes ethical considerations, including acquiring informed consent from participants and preserving participant confidentiality. The text also looks at how studying social capital in Indonesia's urban and rural areas may turn out. (Sponagel et al., 2022). The paper offers a thorough analysis of social capital levels, the various influences on social capital, the consequences of social capital on economic opportunities and general well-being, the current difficulties and potential routes for enhancing social capital, as well as recommendations for policy measures and interventions. (Obstfeld, 2023) This chapter provides a full explanation of the theoretical frameworks, methodology, and prospective conclusions in the study of social capital in Indonesia's urban and rural areas. The determinants of social capital in urban and rural areas of Indonesia can vary due to differences in social, economic, and cultural contexts. Social capital refers to the networks, relationships, and social norms that facilitate cooperation and collective action within a community. There is some variation in the factors that influence social capital in Indonesia's urban and rural areas. The emergence and growth of social capital in metropolitan settings may be seen as a result of a number of variables, including variety, population density, and economic opportunity. The capacity of people from different origins to connect socially and form networks in the setting of varied urban areas may have an impact on the development of social capital. (Eriksson et al., 2019). The creation and expression of social capital can, however, be greatly influenced in rural regions due to a variety of circumstances, including strong familial and community ties, common cultural norms and values, and a lack of resources. The potential influence of kinship networks and the ability of community members to provide support and engage in cooperative efforts may have implications for the development of social capital within rural communities (Jewkes et al., 2023). Trust, reciprocity, social norms, and governance structures are additional factors influencing social capital in urban and rural situations. Learning more about these variables can help in the development of plans and interventions aimed at boosting social capital in Indonesia's urban and rural communities. The theoretical frameworks, research methods, and possible conclusions related to assessing social capital in Indonesia's urban and rural areas are summarized in this section. This study looks at the numerous factors that affect social capital in these particular contexts, such as variety, population density, economic possibilities, strong ties between families and communities, shared values, and scarce resources. Additionally, the text discusses many research techniques, including quantitative, qualitative, and mixed-methods approaches. It emphasizes how important ethical issues are while doing social capital research. (Soler-Gallart & Flecha, 2022). Some of the outcomes of research on social capital in Indonesia's urban and rural areas include an analysis of social capital levels, the identification of social capital-affecting factors, an analysis of the effects of social capital on economic prospects and overall welfare, an analysis of the challenges and opportunities for social capital improvement, and policy recommendations and interventions. This section provides a thorough understanding of the analysis of social capital in a variety of scenarios. (Parker et al., 2022). --- DISCUSSION --- EXAMINATION AND ELUCIDATION OF THE RESULTS A careful examination of the acquired data is needed to analyze and understand the results from the analysis of social capital in both urban and rural areas of Indonesia. Finding noteworthy themes, patterns, and correlations within the data would be the goal of this analysis. (Wang et al., 2021). The research goals and theoretical frameworks used in the study would then be used to develop conclusions. In quantitative data analysis, various statistical techniques can examine the associations between variables and evaluate hypotheses (Camphor et al., 2021). These methods consist of regression analysis, correlation analysis, and descriptive statistics. It would entail examining quantitative data obtained from surveys, questionnaires, or archival sources, followed by statistical analysis to discern trends and patterns in indicators of social capital (Molenberghs et al., 2020). Researchers may get important insights into the amounts of social capital existing in both urban and rural areas, as well as the numerous elements that contribute to its creation and maintenance, through data analysis. Qualitative data can be analyzed using thematic analysis, content analysis, or grounded theory methods to discern significant themes, patterns, and relationships within the data (Lou et al., 2023). This process involves looking at qualitative data from interviews, focus groups, observations, or document analysis. It is followed by a comment that aims to give a complete picture of how people's experiences, beliefs, and actions relate to social capital. By examining the data on how context influences social capital in both urban and rural settings, researchers may gain valuable insight into the difficulties and advantages of social capital in both contexts. The findings should be interpreted within the framework of the relevant scholarly literature and the particular research context. It would entail conducting a comparative analysis of prior research on social capital in Indonesia and assessing how these findings enhance the current body of knowledge and comprehension surrounding social capital (Naranjo-Zolotov et al., 2019). It was imperative for researchers to duly acknowledge and account for the constraints inherent in their study, including but not limited to the sample size and potential biases. Furthermore, they must engage in a comprehensive analysis of the ramifications of their findings on theory, policy, and practice (Abdelgawad et al., 2022) Generally, analyzing and interpreting the findings would entail a methodical examination of the gathered data, identifying significant themes and patterns, and formulating conclusions per the research objectives and theoretical frameworks (Lewis, 2021). It is crucial to base the interpretation of the results on the body of prior research and take the study's limitations into account. The findings of this study have the potential to significantly advance our knowledge of social capital levels in Indonesia's urban and rural areas, its causes, and possible effects on economic prospects and general welfare (Meng et al., 2019). These observations can provide valuable input for developing Prastyo, R., E., Wisadirana, D., Rozuli, A., I., Hakim, M., L. (2024). Social Capital's Impact on Indonesia's Urban and Rural Areas 10 strategies, policies, and interventions to enhance social capital and foster community development in urban and rural settings. --- COMMUNITY DEVELOPMENT AND WELL-BEING AFFECTED BY SOCIAL CAPITAL IN URBAN AND RURAL AREAS The key factors contributing to positive outcomes in a community setting encompass enhanced accessibility to resources, heightened trust and cooperation among community members, and strengthened social support networks (Fielding, 2022). In urban environments, a positive correlation exists between elevated levels of social capital and decreased crime rates, improved educational achievements, and enhanced economic growth. On the other hand, rural regions frequently encounter distinct obstacles, including inadequate infrastructure and geographical seclusion, which can affect the establishment and sustenance of social capital. Understanding the effects of social capital in urban and rural contexts is essential for promoting equitable and sustainable community development. (He et al., 2021). Urban settings with high densities of people encourage the growth of social networks and make it easier for people to share resources and knowledge. The interconnectedness seen here may lead to higher levels of mutual trust, collaboration, and group action within the community (Schmidt et al., 2022). Consequently, urban regions characterized by elevated levels of social capital exhibit a propensity for decreased crime rates as residents actively engage in mutual surveillance and collaborative efforts to tackle community concerns. Furthermore, the correlation between robust social networks in urban regions and improved educational achievements has been established, as it is more probable for parents and community members to promote and endorse high-quality education actively (Muhamad Khair et al., 2020). Furthermore, social capital is paramount in fostering economic development within urban regions. Informal networks and social connections allow individuals to avail themselves of employment prospects, obtain financial assistance, and acquire valuable resources to initiate and expand entrepreneurial endeavors. It facilitates the development of a prosperous regional economy characterized by elevated rates of employment and increased incomes for local inhabitants. In urban settings, the existence of social capital engenders a collective identity. It nurtures collaborative efforts, resulting in many advantageous consequences for individuals and the urban community (Sugawara et al., 2022). Prastyo, R., E., Wisadirana, D., Rozuli, A., I., Hakim, M., L. (2024). Social Capital's Impact on Indonesia's Urban and Rural Areas 11 --- AN EXAMINATION OF THE DYNAMICS OF SOCIAL CAPITAL IN URBAN AND RURAL SETTINGS The analysis of social capital dynamics in urban and rural environments is a field of study that is of great significance. Understanding how social capital operates differently in both environments may provide policymakers and community leaders with important insights. Both settings have unique qualities. Due to the higher population density and wider range of social interaction opportunities, social capital is more diverse and interrelated in metropolitan settings (Zhao et al., 2021). This phenomenon can increase residents' trust and cooperation, fostering greater community involvement and facilitating collective action. Conversely, rural regions frequently exhibit more cohesive communities characterized by the cultivation of social capital through robust interpersonal connections and shared principles (Valentine et al., 2022). Rural areas frequently exhibit the characteristic of tightly knit communities, wherein social capital plays a crucial role in providing support and aid due to potential constraints in resource availability and service accessibility. Social capital is pivotal in cultivating a sense of inclusion and cohesion among individuals within these communities (Galarneau & Smith, 2022). This phenomenon can foster elevated social cohesion and an increased propensity to assist each other during adversity. Moreover, it is worth noting that social capital in rural regions frequently extends beyond proximate neighbors, encompassing extended family members, friends, and even acquaintances within the local community. The interconnectivity and reciprocal support system can enhance the resilience and sustainability of rural communities by establishing a robust basis for collective engagement and cooperative resolution of challenges (Avilez Pineda et al., 2019). During periods of crisis, such as natural calamities or economic recessions, social capital can be crucial in coordinating relief initiatives and offering vital assistance to individuals impacted by the situation. In addition, establishing social capital can cultivate a feeling of inclusion and unity, which can serve as a magnet for individuals to settle and remain in rural regions. This phenomenon contributes to forming a steadfast and interconnected community with enhanced capabilities to confront collective obstacles and collaborate towards mutually beneficial objectives. In rural communities, social capital is paramount in augmenting the overall well-being and quality of life (Jarernpanit, 2022). Prastyo, R., E., Wisadirana, D., Rozuli, A., I., Hakim, M., L. (2024). Social Capital's Impact on Indonesia's Urban and Rural Areas 12 --- CONCLUSION This study underscores the vital role of social capital in rural communities, emphasizing its positive impact on economic development, health outcomes, and community resilience. It calls for intentional resource allocation to cultivate and sustain social capital, contributing to the social capital literature in Indonesia by exploring its influence on economic growth, a less-explored aspect (Cofré-Bravo et al., 2019). The research highlights the importance of understanding the connection between social networks and economic development to inform policy-making for sustainable growth. It urges governments and organizations to recognize and implement policies fostering social capital and promoting benefits such as entrepreneurship, innovation, and knowledge exchange for economic expansion (Dressel et al., 2020). In policy-making, the emphasis is on creating an environment conducive to social capital development through strategies like community programs and mentorship. Policymakers may prioritize formulating policies supporting the establishment of social enterprises, allocating financial resources, and providing necessary support for their expansion (Hannigan et al., 2022).. Policymakers are encouraged to support social enterprises, allocate resources for education, and enhance skills to create a positive feedback loop that empowers individuals, actively contributing to community betterment and broader societal advancement (Kirby-McGregor et al., 2023). Overall, governments are encouraged to recognize the significance of social capital in shaping a promising and economically advantageous future for their constituents. --- JOURNAL OF LAW AND SUSTAINABLE DEVELOPMENT Prastyo, R., E., Wisadirana, D., Rozuli, A., I., Hakim, M., L. (2024). Social Capital's Impact on Indonesia's Urban and Rural Areas JOURNAL OF LAW AND SUSTAINABLE DEVELOPMENT Prastyo, R., E., Wisadirana, D., Rozuli, A., I., Hakim, M., L. (2024). Social Capital's Impact on Indonesia's Urban and Rural Areas JOURNAL OF LAW AND SUSTAINABLE DEVELOPMENT Prastyo, R., E., Wisadirana, D., Rozuli, A., I., Hakim, M., L. (2024) Bendel, D., Angenendt, E., Weber, T. K. D., Gayler, S., Streck, T., & Bahrs, E. (2022). Integrated assessment of regional approaches for biodiversity offsetting in urban-rural areas -A future based case study from Germany using arable land as an example. Land Use Policy, 117, 106085. https://doi.org/10.1016/J.LANDUSEPOL.2022.106085 JOURNAL OF LAW AND SUSTAINABLE DEVELOPMENT Prastyo, R., E., Wisadirana, D., Rozuli, A., I., Hakim, M., L. (2024)

Objective: An intangible resource called social capital is essential for promoting social cohesion, advancing the economy, and improving the general welfare of people and communities. It may be divided into two categories: bridging social capital, which deals with interpersonal connections between various groups or communities, and bonding social capital, which is concerned with interpersonal connections inside a certain group or community.The approach utilized in this article is a literature review, a method of gathering information that entails comprehending and exploring theories found in different literature sources relevant to the research.In order to comprehend how social capital affects social cohesion and community development, as well as the similarities and variations between urban and rural communities in Indonesia, it is important to first understand how social capital is formed and maintained.The research assesses the challenges and possible benefits related to increasing social capital in Indonesia's urban and rural locations.

Introduction Access to sexual and reproductive health (SRH) services by adolescents and youths has been a serious public health concern globally since the United Nations (UN) coordinated an international Conference in Cairo in 1994 [1]. Subsequently, the Colombo Declaration on Youth [2] and the UN Sustainable Development Goals (SDGs) 2030 Agenda, [3] recommended universal access to SRH services and information for youths with the pledge of "leaving no one behind" [3]. These are core requirements to ensure that young people, especially young women, can make informed choices about when to have children and prevention of sexually transmitted diseases [4]. Despite these global commitments and targets, there is a dearth of literature on the factors associated with access to SRH services in government health facilities by immigrant youths in a high-migrant populated community like Hillbrow. In South Africa, immigrant youths aged 18-34 accounted for 29.4% of all international migrants and 9.2% of all youths from 18 to 34 years in 2019 [5]. Hillbrow is an immigrant-dense urban suburb of Johannesburg, with a total population of approximately 74,000; of which 58% are between 18 and 34 years, and 27% are immigrant youths [6]. Researchers globally have examined factors influencing knowledge of family planning KFP and access to SRH services by immigrant youths [7,8]), which include gender, age, level of education, wealth index [9] and migration status [10,11]. Other influencing factors identified by researchers include experiences of discrimination and lack of social support [12,13]; as well as attitude of health care practitioners [8,14]. Increased international migration, as well as the increasing number of undocumented migrants, has heightened concerns about immigrants being exposed to health risks. It is claimed that the South African laws, policies, and guidelines "are progressive and comprehensive on contraceptive service provision in the public sector; and promote integrated, rights-based service delivery" [15]. Negative consequences of poor SRH outcomes-such as unplanned and unwanted pregnancies, HIV/AIDS, sexually transmitted infections (STIs), maternal mortality and infertility-these all affect the development indicators of the country. In order to curb the negative SRH outcomes among the entire population, South Africa is committed to the SDGs, especially SDG 3.7 on universal access to SRH services. Access to health facilities for SRH services is considered vital for all youths. Since sexually active youths interact within the same social space, it is important for all youths to have access to SRH services, irrespective of migration status. --- Plain language summary Access to SRH services by adolescents and youths has been a serious public health concern globally. In South Africa, universal access to health care including reproductive health care services is contained in Sect. 27 of the Constitution. The country is committed to the 2030 Agenda of the SDGs with the theme "leaving no one behind". The objective of this study was to examine the factors associated with knowledge about family planning and access to SRH services among sexually active immigrant youths in Hillbrow, South Africa. Interviewer-administered questionnaires were used to collect information on the socio-demographic, migration, and reproductive health characteristics from 439 sexually active immigrant youths in Hillbrow during December 2019. The data were analysed to assess the factors associated with knowledge about family planning and access to SRH services. The findings show that 46% of the immigrant youth had poor knowledge about family planning, while 35% had no access to SRH services, 42% had some access and only 23% had access. The study showed that the factors associated with knowledge about family planning are female respondents, from the rich income group, those in the older age category, those who received information about family planning 6-months prior to the survey and those using a contraceptive at the time of the survey. The factors associated with having no access to SRH services were those respondents having completed secondary education and higher, from the rich income group, undocumented immigrants, those respondents having experienced discrimination in Hillbrow, youths not having received information about family planning prior to the survey, and those not using a contraceptive. It is important to make access to SRH services available to immigrant youths to prevent associated negative SRH outcomes among the youths. Despite the growing population of immigrant youthsboth documented and undocumented, in South Africa, specifically in Hillbrow with the highest population of immigrants in Johannesburg, little is known about their knowledge about family planning. Furthermore, there is scanty of evidence about their access to SRH services from government health facilities in their neighbourhood (Hillbrow Community Health Centre and General Clinic, Hillbrow). Based on the Behavioural Model of Health Service Utilization (BMHSU) and the modified model on health service utilization among immigrants, [16,17], the main objective of this study was to determine the factors associated with KFP and access to SRH services in government health facilities by sexually active immigrant youths in Hillbrow, South Africa. --- Methods This was a cross-sectional study involving a household survey that used interviewer-administered questionnaires to gather information from immigrant youths (18-34 years), living in Hillbrow, in December 2019. Data on 437 sexually active respondents was analysed in STATA 14 using univariate, bivariate, logistic, and multinomial regression models. A p-value of <unk> 0.05 was chosen as the level of significance. The present study is part of a larger study that examined the "Sexual behaviour of immigrant youths and their access to reproductive health services in Hillbrow, South Africa". The study population consisted of immigrant youths only, both male and female, 18-34 years of age, who had crossed an international border into South Africa, both documented and undocumented. Data on socio-demographic characteristics, migration, and sexual and reproductive health were collected from consenting immigrant youths. The sample size for frequency in a population with a correction factor was used to calculate the sample size of immigrant youths needed for this study to account for the effect of intra-cluster correlation or the design effect [18,19]. The Open Epi Version 3.05.07 software for estimating a population proportion with specified absolute precision was used to calculate the sample size. The design effect (DEFF) of 1.25 was used in order to correct for estimated sample variance and account for within group differences among the immigrant youths. The following formula for sample size calculation was used: where, n = sample size, N = Population size of immigrant youth in Hillbrow, p = Probability of exposure in the population, d = level of precision margin of error, (1) n = DEFF * Np(1 -p) d 2 z 2 1-<unk> 2 * (N -1) + p * (1 -p) deff = design effect = 1.25, Z = the selected critical value of desired confidence level at 1.96 for the 95% CI. Using this equation, 417 samples were required for this study to detect a difference as statistically significant at 5% level of significance assuming a design effect of 1.25. A nonresponse rate of 12% was factored into this study. This was because of the nature of the primary study that asked questions on sexual behaviours and some undocumented immigrants may be reluctant to participate in the study. This increased the sample size to 467. The design effect of 1.25 was used to correct for estimated sample variance and account for within group differences among the respondents. The study used a two-stage cluster sampling technique due to lack of a sampling frame and cost. The recruitment of subjects involved the selection of nine (9) enumeration areas (EAs) were made of the 30 EAs in Hillbrow. Figure 1 presents a flow chart of the two-stage sampling technique. A list of all the streets in the selected EAs was mapped, to form the sampling frame of the clusters. A sample of the clusters (streets) was selected as the first stage sampling. From each of the selected clusters (streets), a sample of the specified number of units (households) were selected as the second stage sampling, and the households were those included in the survey. One respondent per household was included in the survey. Questionnaires were administered to subjects who met the inclusion criteria for the study. The inclusion criteria for respondents were immigrants, between 18-34 years of age, resident in Hillbrow with the ability to provide informed consent. Exclusion criteria were internal migrants, non-migrants, migrants not residing in Hillbrow, anyone who is mentally ill or drunk. The study followed the assumptions of two-stage clusters sampling of each cluster being internally heterogenous and homogenous to all the other clusters, non-overlapping, and good representation of the study population. The questionnaire was structured into themes to capture the different characteristics such as socio-demographic, sexual behaviour, knowledge, and information about SRH and migration characteristics of the respondents. Most of the questions were adapted from the survey questionnaire by the African Population and Health Research Centre (APHRC) and the 2016 South African Demographic and Health Survey [20,21], after an extensive literature review. The questionnaire was pilot tested and revised before the field work was undertaken to collect the data. The outcome variables measured were KFP and access to SRH services from health facility. Table 1a presents the questions and distribution of responses used for generating the outcome variable, KFP. Using principal component analysis (PCA), a composite variable for KFP was generated from two questions on having heard of a family planning method and knowledge of place to obtain contraceptives, each with a list of contraceptive methods. The Cronbach's alpha for the two variables were 0.83 and 0.86 respectively. The new variables generated were "Heard of contraceptive method" and "Know place to obtain a contraceptive method". These two variables were used to generate the variable "knowledge of family planning" (KFP) with "inadequate knowledge" or "adequate knowledge" categories. For the second outcome variable, access to health facility for SRH services, respondents were asked about the challenges they have encountered in using the SRH services in the government health facilities in their community. There are two government health facilities in the community-the Community Health Clinic and the municipal clinic, popularly called General clinic. The challenges listed included high cost, discrimination, attitude of health facility staff, waiting time, language barrier, distance, fear and safety concerns, not knowing about services, thinking they do not need services and lack of legal document (immigration permit to be in South Africa). The summary is presented in Table 1b. A composite variable for access to health facility for SRH services was generated using PCA. The new variable-access to health services-has three categories: no access, some access and access. The selection of explanatory variables was based on the conceptual framework for this study, which proposes the influence of demographic factors (age, gender, marital status); socio-economic factors (wealth index, level of education attained); migration and community social factors (migration status, discrimination and xenophobia, social support), and information about SRH (having recently received family planning messages and use of family planning at the time of the study), on access to health facility for SRH services by immigrant youths. The distribution of the outcome variables is shown in Table 1c. Data analysis was conducted using STATA MP 14. Data was analysed at three levels-univariate, bivariate and multivariate analyses. The univariate analysis used descriptive statistics, percentages, and frequencies to describe background characteristics including demographic, socioeconomic, migration and SRH factors of the respondents. The second level of analysis, the bivariate analysis, assessed the associations between the explanatory variables and the two outcome variables using Pearson's chi-square test. Explanatory variables were considered significant at a p-value of 0.05 or less and 95% confidence interval. The third level of analysis, the multivariate analysis, was conducted to assess the adjusted association of the explanatory variables with KFP and access to SRH services respectively. A multivariate logistic regression model using adjusted odds ratio (AOR) was used to determine the factors associated with KFP in the presence of the explanatory variables. A binary logistic regression model was used because the outcome variable KFP is dichotomous (inadequate knowledge and adequate knowledge), while a multinomial logistic regression model was used for the outcome variable, access to SRH services, which has three categories (no access, some access and access). --- Results The selected background characteristics of 439 sexually active immigrant youths are summarised in 0.17, CI = 0.10 -0.29). Similarly, immigrant youths who were not using a family planning method at the time of the survey had reduced likelihood of having knowledge about family planning (AOR: 0.37, CI = 0.19-0.70). --- Factors associated with access to SRH services from government health facilities by sexually active immigrant youth The unadjusted multinomial logistic regression for access to SRH services and socio-demographic characteristics is shown in Table 5a. With access to SRH services as the base outcome, the socio-demographic factors associated with no access and some access were highest level of education attained and wealth quintile. The bivariate result showed that immigrant youths with complete secondary education and higher have increased risk to have no access to SRH services in government health facilities (RRR = 2.64, 95% CI = 1.55-4.48, p <unk> 0.001). Similarly, the relative risk for having some access to SRH services for those with secondary education and higher was increased by a factor of 2.55 (RRR = 2.55, 95% CI = 1.53-4.25, p <unk> 0.001). There was an increased relative risk of having no access to SRH services for immigrant youths in rich wealth quintile (RRR = 3.48, 95% CI = 1.67-7.23, p-value <unk> 0.001). Similarly, there was an increased relative risk of having some access by those in the middle (RRR = 2.10, 95% CI = 1.22-3.61, p-value <unk> 0.01) and rich wealth quintiles (RRR = 3.0, 95% CI = 1.42-6.30, p-value <unk> 0.001). Table 5b presented the unadjusted multinomial regression of migration and reproductive health characteristics. With access to SRH services as the base outcome, the migration and reproductive health factors associated with no access were migration status, having received information about family planning 6 months prior to the survey and use of contraceptive methods. The factors associated with having some access were migration status, duration of stay in Hillbrow, having social support in in South Africa, having received information about family planning 6 months prior to the survey and use of contraceptive methods. The relative risk of having no access by undocumented immigrant youths decreased by a factor of 0.37 compared to documented immigrants (RRR = 0.37, 95% CI = 0.22-0.62, p-value <unk> 0.001). Similarly, among those with some access to SRH services, the relative risk of undocumented immigrant youths decreased by a factor of 0.36 (RRR = 0.36, 95% CI = 0.21-0.60, p-value <unk> 0.001). Immigrant youths who have lived in Hillbrow for 3-5 years have increased relative risk for some access to SRH services compared with those with 0-2 years of stay (RRR = 2.35, 95% CI = 1.24-4.45, p-value <unk> 0.01). Similarly, those who have lived for 6-20 years in the community have increased relative risk of having some access (RRR = 1.97, 95% CI = 1.04-3.73, p-value <unk> 0.05). The relative risk of having no access by immigrant youths who did not receive information about family planning 6 months prior to the survey decreased by a factor of 0.48 compared to those who received such information (RRR = 0.48, 95% CI = 0.28-0.80, p-value <unk> 0.01). Similarly, the relative risk of having no access to SRH services by immigrant youths who were not using a contraceptive method at the time of the survey decreased by a factor of 0.56 compared to those who were using contraceptives at the time of the survey (RRR = 0.56, 95% CI = 0.31-0.99, p-value <unk> 0.05). The results of adjusted relative risk ratio on factors associated with access to SRH services by immigrant youths is presented in Table 6. With access to SRH services as the base outcome, the factors associated with no access to SRH services were level of education attained, wealth quintile, migration status, having experienced discrimination in Hillbrow and having received information about family planning 6 months prior to the survey. The results showed that immigrant youths with completed secondary education and higher have increased relative risk to have no access to SRH services compared to their counterparts with primary or incomplete secondary education (ARRR = 1.89, 95% CI = 1.06-3.36, p-value <unk> 0.05). Compared with those in the poor wealth quintile, those in the rich quintile had an increased relative risk of having no access to SRH services (ARRR = 2.25, 95% CI = 1.00-5.07, p-value <unk> 0.05). Compared to the documented, undocumented immigrant youths had a reduced risk of having no access to SRH services (ARRR = 0.49, 95% CI = 0.27-0.88, p-value <unk> 0.05). Immigrant youths who reported not having experienced discrimination in Hillbrow were at increased risk for not having access to SRH services (ARRR = 2.06, 95% CI = 1.15-3.67, p-value <unk> 0.05). Compared to immigrant youths who received information about family planning 6 months prior to the survey, those who did not receive such information had reduced risk of no access to SRH services (ARRR = 0.49, 95% CI = 0.26-0.90, p-value <unk> 0.05). Using access as the baseline outcome, the determinants of having some access to SRH services by immigrant youths are being undocumented (ARRR = 0.51, 95% CI = 0.29-0.91, p-value <unk> 0.05), not having received SRH information 6 months prior to the survey (ARRR = 0.38, 95% CI = 0.21-0.70, p-value <unk> 0.01) and not using any family planning methods at the time of the survey (ARRR = 0.31, 95% CI = 0.16-0.62, p-value <unk> 0.001). --- Discussion This study examined the determinants of KFP and access to SRH services provided by government health facilities among sexually active immigrant youths in Hillbrow, South Africa. The study identified inadequate KFP and lack of access and some access to SRH services from government health facilities in Hillbrow as factors contributing to negative SRH outcomes. It further showed that majority of the participating immigrant youth's sourced information about SRH issues from unreliable sources. They did not get information about SRH from the health facilities in their neighbourhood though they are within walking distance from their residence. This finding is of concern because of the social environment of Hillbrow, which consist of a mix of internal migrants, non-migrants, and international migrants; with lots of social activities including risky sexual behaviours and alcohol consumption [22]. Lack of access to SRH services including correct, reliable information about SRH issues resulted in poor knowledge about family planning and eventually negative SRH outcomes including STIs, re-infections, HIV infections, unplanned and unwanted pregnancies, unsafe abortions, and maternal mortality in Hillbrow, because sexual relationships among youths in Hillbrow do not respect migration status. The observed gender differences, with more males than females lacking knowledge about family planning, may be a consequence of more females seeking SRH information and services due to pregnancy and childbearing needs. Participants in the group with a higher wealth index have better knowledge, but they have no access to government health facilities. This indicates that the rich acquire SRH services, including knowledge about family planning, from sources other than government health facilities. Despite the availability of youth friendly SRH services for adolescents and youth under 25 years of age in Hillbrow, this study shows that immigrant youth aged 18-24 have poor knowledge about family planning. This study further shows that those immigrant youths without documentation and those who have experienced perceived discrimination had no access to SRH services from government health facilities. Qualitative study among the study population will provide a better understanding. The association between not having received any information about family planning six months prior to the survey with poor knowledge about family planning, having no access or having some access to SRH services underscore the importance of education, information, and communication about family planning among immigrants in Hillbrow. A previous study in South Africa that used a cross-sectional design in migrant-dense areas recorded poor information about SRH services and products among immigrants [20]. Findings of lack of access to SRH services in by immigrants is consistent with a study on pregnant women's attendance at three urban antenatal care (ANC) settings in Johannesburg, which showed that the Hillbrow health facility had the lowest levels of ANC attendance, HIV testing and the highest HIV prevalence of all three facilities. The study concluded that the low levels recorded could be attributed to perceived or real lack of access to the services by immigrants [23]. The finding that immigrant youths who were not using contraceptives lacked KFP showed the importance of acquiring knowledge preceding use of contraceptive methods. This is consistent with findings by researchers in Sweden that showed that current use of family planning was associated with KFP [14]. Similarly, Bersamin and colleagues [24] found poor knowledge about accessing SRH services among young men. Lack of access to government health facilities left immigrant youth in Hillbrow with no other choice other than to obtain SRH information from unreliable sources. Qualitative studies are necessary to understand the reason for widespread lack of access to government health facilities by immigrant youth in Hillbrow. The present study should be interpreted with regards to some important limitations. Firstly, it is a cross-sectional study, thus conclusions cannot be drawn on causal relationships. The techniques used cannot predict immigrant youth's knowledge about family planning and their access to SRH services over a period of time. The selection of the study site was not random, but was purposefully selected based on the prolific number of immigrants in the locality. Additionally, the study did not do a verification of immigration status of the respondents. Some respondents may not have been truthful about their immigration status for fear of being apprehended. This could have resulted in an under-reporting of migration status by undocumented migrants. However, the use of well-trained field workers and gatekeepers that assured respondents that the survey was purely for academic purposes likely mitigated this bias. Recall bias might be a limitation in participants reporting having received information about sexual and reproductive health 6 months prior to the survey. Another limitation is that the outcome variable considered access to SRH services from government health facilities which included the Hillbrow community clinic, and the General clinic (a municipal clinic) did not include the private clinics in Hillbrow. Despite these limitations, the study did however collect quantitative data and has shown the factors associated with knowledge about family planning and access to SRH --- Conclusion Despite the perspective of a progressive approach for rights-based service delivery proposed by South Africa's laws, policies and guidelines on contraceptive service provision in the public sector, this study has shown that in practice the immigrant youths living in Hillbrow lack the necessary access to SRH services in government health facilities. Lack of correct information, knowledge and access to SRH services from health facilities by immigrant youths raises questions about implementation of the policies and guidelines for universal access to SRH services in government health facilities. Considering the high population of youths in Hillbrow and the level of social interactions, including risky sexual behaviours between immigrants and non-immigrants, it is necessary to provide access to SRH services to immigrant youths, in order to prevent re-infections of native-born youths. There is need therefore to advocate for the universal access to SRH services provided by government health facilities to curb the negative SRH outcomes, including unplanned pregnancies, unsafe abortions, HIV infections and re-infections among immigrants and non-immigrant youths, in a social space such as Hillbrow; to enable South Africa to effectively make progress to meet the SDG health targets, especially SDG 3.7 and 5.6, relating to SRH services [3,4]. --- Availability of data and materials The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request. --- Variable Base --- Declarations Ethics approval and consent to participate Signed informed consent was obtained from all the study participants prior to the commencement of the interviews. This study was approved by the Human and Social Sciences Research Ethics Committee of the North-West University, South Africa: NWU-01909-19-S7. This manuscript does not contain any individual person's data in any form. --- Competing interests The authors declare no competing interests. • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year --- • At BMC, research is always in progress. --- Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research? Choose BMC and benefit from:? Choose BMC and benefit from: --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: In South Africa, universal access to health care services, including those relating to sexual and reproductive health (SRH) care, is contained in Section 27 of the Constitution and commits the country to supporting the United Nations 2030 Agenda for the Sustainable Development Goals (SDGs). The objective of this study was to examine the factors associated with knowledge about family planning and access to SRH services among sexually active immigrant youths in Hillbrow, South Africa.This cross-sectional study was based on data from a household survey conducted in Hillbrow during December 2019. Interviewer-administered questionnaires were used to collect information from immigrant youths (18-34 years old). Data on 437 sexually active respondents was analysed in STATA 14 using univariate, bivariate, logistic, and multinomial regression models. A p-value of < 0.05 was chosen as the level of significance. Results: About half of the respondents had poor knowledge about family planning; about one-third (35%) of the immigrant youths had no access to SRH services, 42% had some access, and 23% had access. The adjusted logistic regression model showed that being a female (AOR = 3.85, CI: 2.34-6.35, belonging to age group 30-34 years (AOR = 3.88, CI: 2.00-7.53); belonging to the rich wealth index (AOR = 2. 55 (1.32-4.93); not having received information about family planning (AOR = 0.17, CI = 0.10-0.29) and not using a contraceptive at the time of the survey (AOR = 0.37, CI: 0.19-0.70) were factors associated with having knowledge about family planning. The adjusted multinomial regression shows that the factors associated with not having access to SRH services were secondary or higher level of education (ARRR = 1.89, 95% CI = 1.06-3.36), belonging to the rich wealth quintile (ARRR = 2.25, 95% CI = 1.00-5.07), being undocumented (ARRR = 0.49, 95% CI = 0.27-0.88), having experienced discrimination in Hillbrow (ARRR = 2.06, 95% CI = 1.15-3.67) and having received information about family planning 6 months prior to the survey (ARRR = 0.49, 95% CI = 0.26-0.90, p-value < 0.05).

Background Symptoms of mental health and substance use (MHSU) challenges start early in life, with 50-70% developing during childhood or adolescence [1]. The peak age of onset for MHSU disorders is 14.5 years (k = 14, median = 18, interquartile range (IQR) = 11-34) [2]. In Canada, 20% of youth have a mental disorder [1], and one in four young people (ages 12-24) experience a MHSU disorder each year, with most presenting to the emergency department (ED) as their first point of contact for health services [3,4]. According to recent surveillance data from the Canadian Institute for Health Information (CIHI) between 2008 to 2019, there was a 61% rise in mental health related ED visits by Canadian children and youth [1]. Evidence also suggests that more than half of Canadian youth and young adults presenting to the ED with MHSU needs have not had any previous mental health related-contact with the health-care system [5,6]. In BC, the opioid use epidemic, declared a public health emergency in 2016, has compounded the MHSU concerns for young people, with youth and young adults accounting for 20-25% of all opioid-related overdoses and deaths [7]. In two other Canadian provinces (Ontario and Alberta), young people aged 15-24 had the highest rates of ED visits for opioid poisoning from 2010 to 2011 and 2014-2015 [8]. Exacerbating these trends, there is currently a profound disconnect between the ED and MHSU services for young people. Young people are presenting to the ED as their first point of contact for MHSU concerns but due to various reasons may not be admitted to the hospital [4]. Thus, it is necessary that ED staff are equipped to refer and connect such young people upon presentation to the appropriate MHSU services for care and follow-up. Integrated youth services (IYS) are one example of a community-based, youth-and family-centred MHSU service. Integrated youth services provide multidisciplinary care (including mental, physical and social aspects), and are preferably in one location that is a youth-friendly environment [9,10]. Young people and their families and needs are prioritized in design of services and care [11]. IYS exist internationally, and Hetrick and colleagues provide a detailed overview of development and evaluation from a global perspective [9]. Examples of IYS internationally include headspace TM in Australia [12], Jigsaw in Ireland [13], Irish Youth One Stop Shops (YOSS) [14], Les Maisons des Adolescents in France [15], and Youth One Stop Shops (YOSS) in New Zealand [16]. IYS have also been established in Canada; however, they are more infant in their development: Youth Wellness Hubs Ontario (YWHO) [17], and Foundry [11,17]. IYS have existed for over two decades (the earliest IYS was established in New Zealand in 1993) [16]. Pathways to such services (and other mental health care for young people) are complex, involve multiple contacts and can have delays to getting appropriate care [18]. Important areas of care pathways to consider include primary care and the roles of families. In addition, the high rate of emergency services involvement in such pathways makes the care pathway between the ED and mental health services for young people necessary to consider [18]. Little is known about referral pathways for young people between the ED and IYS offering MHSU support. There is an immediate need to improve referrals to organizations that can provide ongoing services for MHSU concerns in order to deliver appropriate care and prevent repeat ED visits and hospitalization for MHSU disorders. There were two main objectives of this study: (1) to better understand the assessment, treatment, and referral for young people (ages 12-24 years) presenting to the ED with MHSU concerns; and (2) to explore how to improve the transition from the ED to IYS for young people with MHSU concerns. --- Methods --- Setting This study took place in British Columbia (BC), Canada. In the province of BC, Foundry is an integrated youth service (IYS) in the community that provides care (including MHSU services) for young people ages 12-24 years [11]. There are currently eleven centres across BC, with twelve more in development, and a provincial virtual care service [19]. Foundry has five core service streams, including mental health, substance use, peer support, primary care, and social services. --- Approach Since little is known about the transition from the ED to IYS for young people with MHSU concerns in Canada, we conducted a qualitative study for exploratory purposes. The Standards for Reporting Qualitative Research (SRQR) was referenced to guide study reporting [20]. --- Participants Ethical approval was obtained from the University of British Columbia Ethics Board (H20-01396). To be eligible for participation, individuals had to live in the province of BC and self-identify as one of three stakeholder groups: (1) health care providers with experience in the ED (e.g., nurses, physicians) or experience working with individuals with MHSU disorders in any setting; (2) individuals with experience working for an agency or institution in any capacity focused on young people with MHSU concerns; and/or IYS (3) young people (age [16][17][18][19][20][21][22][23][24] and/ or family members of young people with personal lived experience with MHSU concerns. Participants were recruited from across the province of BC, Canada, through purposive snowball sampling. The authors FS DB and SM are health care providers working in ED and IYS settings. All reached out to colleagues that met eligibility requirements to begin this process. The aim was to recruit diverse stakeholders from a variety of communities considering population centre density, their local services, and communities with and without IYS in the community. --- Data collection and procedures Semi-structured one-on-one video (over the platform Zoom Video Corporation, San Jose, CA) or telephone interviews were conducted by MW, JW, SB, and FS. MW and JW were research assistants at the time of the study. Data collection occurred June-August 2020 with rolling admission into the study. Participants provided verbal or written consent at the beginning of the interview after reviewing an information sheet about the study. As part of this process, participants were informed that all questions were optional and that they could end the interview at any time. The semi-structured interview (Table 1) addressed the care experiences of young people accessing the ED for MHSU concerns, including assessments used and treatments provided. Questions also focused on the transition and referral pathways from the ED to IYS in the community such as Foundry, and vice versa. Lastly, considerations were given to what an ideal process would look like for the transition between the ED and services in the community, as well as improving young people's experiences while in the ED. Following the interview, participants were asked five optional demographic questions about age, gender, race, role/position, years in practice, and community of practice as applicable. Participants were provided a $40 gift card for participating. --- Data analysis and theory All interviews were audio-recorded and transcribed verbatim. To maintain confidentiality, unique participant numbers were assigned and all potentially identifying information was removed from the transcripts. Analysis was conceptually guided by the Social Ecological Model (SEM) [21]. The SEM is a framework that helps to understand how individuals interact with their environment within a complex social system and considers various connections to an individual [21]. It considers and acknowledges that important determinants to health exist at multiple levels, including that of the individual, interpersonal, organizational, community and policy (Fig. 1). The SEM has been used previously and acknowledged as helpful for understanding suicide assessment and prevention [22], for transitions of young people within health care services [23], and acknowledged as important in considering experiences in an ED after opioid overdose [24]. The SEM has also been recommended when considering transformation of traditional silos and acute treatment in health care, acknowledging it is necessary to consider social determinants of health for individuals [25]. The research team considered the individual a young person with MHSU concerns and determined definitions of all levels of the SEM relevant to the research objectives to guide analysis (Table 2). A deductive and inductive analytical approach was used in a multi-step process to conduct a thematic analysis [26,27]. NVivo 12 Software (QSR International, Doncaster, AU) was used to manage data and support qualitative data analysis. KG & MW engaged in active analysis through several phases to ensure thoroughness [27]. The phases included becoming familiar with the data-listening to audio recordings and reading transcripts while notes were taken. Next, codes were generated in NVivo based on the coding framework created using the SEM and definitions. This step was considered deductive, and Organizational level The emergency department within the hospital. Examples of factors considered were rules and regulations; hours; resources such as beds, staffing, etc. While focus was on the ED, the larger institution of the hospital and the ED within it were also considered. Community level IYS services in the community such as Foundry, and other services for young people with MHSU concerns. Factors here included the transition from the ED/hospital, access to the community services (post-discharge from the ED), as well as the care pathway from the community back to the ED. --- Policy level Local, provincial, and national laws/regulations relevant to the health and health care of a young person with MHSU concerns. Examples of factors considered were charting systems throughout the province, bills, provincial health coverage. coding was conducted by KG & MW independently. Regular meetings were held to ensure understanding of the codes and concept definitions and KG and MW coded one transcript together during this time. The broader codes were then reviewed to search for and identify themes inductively. This entailed searching through each level of the SEM codes to find data-driven ideas that were commonly occurring relevant to the research question (and ultimately developed into themes). Themes were then reviewed across and within the entire data set (and within and across SEM levels) and discussed by KG & MW to stimulate dialogue and encourage researcher reflexive acknowledgment of results and perspectives in the research process [28]. Themes were then defined and named. Writing of themes was an iterative and an integrated process throughout to aim for a comprehensive analysis. To aim for quality thematic analysis, the 15-point "checklist" for good thematic analysis was consulted [26]. --- Findings --- Characteristics of study participants The sample included 26 individuals with a mean age (<unk>SD) of 42 (<unk>12.3). A variety of stakeholders participated including ED physicians (n = 6), social workers (n = 4), nurses (n = 2), an occupational therapist (n = 1); a counselor (n = 1); staff/leadership in IYS organizations (n = 4); mental health/family workers (n = 3); peer support workers (n = 2), and parents (n = 3). For service providers, the average number of years in practice (<unk>SD) was 14 (<unk>7.4). The majority identified as women (19/26; 73%) and white/Caucasian (20/26; 78%). Participants were from 9 different communities with over half (n = 15) being from a medium sized urban population centre as defined by Statistics Canada [29]. Further details of participant characteristics are provided in Table 3. Demographic questions were open ended and so responses to questions vary (e.g., race/ethnicity identity). --- Findings of the thematic analysis The themes and sub-themes are discussed in detail below and outlined in Fig. 2 in relation to the Social Ecological Model. --- Theme 1 Context of the Environment Participants discussed the impact that the ED and hospital contexts have on young people's transitions from ED to IYS, with a focus on staff schedules, hours of operation, wait times, and the physical environment. --- Environment of the Emergency Department and Hospital Within the ED, participants discussed that time played a significant role in the experience of a young person presenting with MHSU concerns. This was in regard to the timing of MHSU designated staff shifts in the ED, the amount of time waiting and how it was spent while waiting in the ED to be assessed, admitted or discharged. Participants expressed concern for how long a young person waits in the ED when they are in a crisis and presenting in distress. For instance, participants spoke of this experience taking up 12 hours or days, and that health Another participant expressed a similar concern and said, "There's often long wait periods, historically I've experienced with young people where they've been left in ER hallways for hours and sometimes days. " (Early Intervention Assertive Case Manager, Participant #15). In addition to wait times, participants spoke of the timing of staff shifts in the ED, which could hinder or enhance a young person's experience. For example, when young people present to the ED "after hours" there are usually less staff, resulting in higher wait times. Participants also spoke of specialized MHSU staff in the ED, including child psychiatrists and registered nurses with an addiction or mental health specialty. Although these service providers were seen to enhance care, participants noted that their hours of work were limited (e.g., Monday-Friday during the daytime) and so weren't always available when needed. These aspects of time and resources (wait times and staff shifts) were considered at the organization level of the SEM. A parent highlighted this issue: "The really big problem is that often he [young person] would go to the Emergency Department at the end of the week and then these, the addiction team isn't available on the weekends or the evenings so instead of keeping him over the weekend so he could see someone, they just discharged him onto the street, like he can't stay with us because of his issues and because of my daughter so he just goes back onto the street, it's pretty sad. " Parent, Participant #17. Another participant acknowledged the limited resources in the ED and their experience: "When I was a young person my experience in the ER was a life changing one for me but had more to do with the other lived experience in the ER than it did for any of the staff that were there. I think I could really tell that the staff were super burnt out, like they didn't have the resources that they want and that kind of left them not able to listen or hear the stories of people who were there with them but for me the like being able to connect with other people who have lived experience that was there really life changing thing. " IYS Staff, Participant #4. --- Fig. 2 Results of thematic analysis and the Social Ecological Model Participants also spoke about the physical environment of hospital EDs and the impact this could have on young people. Within the ED, participants expressed concern over poor lighting, lack of private spaces or rooms while waiting, and congestion of people making it "quite loud and busy" (Early Intervention Assertive Case Manager, Participant 15). These were considered organization level factors of the SEM. Participants also discussed that there are a limited number of beds within the ED and for admission to a certain ward (psychiatric wards and pediatric wards were discussed). In addition, depending on age or certain MHSU needs, admission to either of these units may be inappropriate. Participants also spoke of a lack of youth-oriented substance use treatments available (e.g. detox beds, pharmacological and psychosocial treatments) in a hospital and in the community. Bed shortages and units/wards were considered policy level factors of the SEM. One participant discussed both the environment in the ED and admission to a unit: "You've got multiple nurses coming and going in the next thing you know security guards are coming in and all kinds of stuff so you know you're just kind of whatever you're just sort of pushed off to the side, they don't have the capacity to deal with it and it's not a nice patient to deal with right?... By the time they stabilized, they're ready to get out of there because they are in a pediatrics unit so it's already odd because you know they're next to a child who may have a broken leg or you know a baby that's in there for a broken arm, it's a mixed bag. There really isn't any place for them". IYS Leadership, Participant #23. --- Environment of Integrated Youth Services Participants were concerned regarding the hours of IYS availability and access to specialists (such as a psychiatrist). Generally, community IYS are open from 8:30 am-4:30 pm on weekdays; if a young person is discharged from the ED at any other time, they are left to initiate IYS contact themselves. These were considered community level factors of the SEM. One participant Fig. 3 A Parent's story of waiting with their young person at the ED.*. *Note: All participants received the interview questions ahead of time and this participant came prepared with a written script and read from the script. They also then provided the written script to the research team discussed the repercussions of the limited IYS hours and connection to ED discharges: "This young person just showed up and we told them to come to Foundry, if there had been more ongoing care then there might be some discharge planning and some connections that are available and built but kind of quite often young people and their families are just left to navigate that in some ways on their own... you know I know it's more challenging that we're [IYS] kind of a Monday-Friday 9-5 industry and they're a 24-hours a day, 7 days a week kind of industry, right? So how do we do that on a Saturday evening after a young person has overdosed or maybe showed up because you know they're at risk of harm... This really came up for us over the last year and a half, we've had, we had sort of a period of you know 3 young people who had died by suicide and then within a couple months another couple people you know just really realizing that you know sort of those bigger players that interact with young people like mental health and substance use services, education services and others really don't know what's happening on the weekends, right?" IYS Leadership, Participant #7. --- Theme 2 Assessments Participants that worked in the ED spoke of not knowing how to assess for MHSU concerns and using MHSU-related assessments, especially for young people. Standardized scoring systems and needing to get a certain score to show a level of crisis through reported symptoms were also discussed. Lastly, participants discussed current assessments used and recommendations for improvement. All factors about assessments in the ED were considered the organizational level of the SEM. --- Not for young people Participants were unclear with regard to youth-appropriate ED assessments and resources and described a lack of confidence and knowledge when assessing or referring young people with MHSU challenges. In response to being asked about assessments a physician responded: "Well there are certainly different scoring systems that we use, scoring system that we use for alcohol intoxication, COWS [The Clinical Opiate Withdrawal Scale] is a scoring system that we'll use for alcohol overdose or opioid withdrawal symptoms. I know that there are several agitation scores that do exist, we don't often or certainly I don't use those I don't have a numeric scoring system that I use for agitation but I do know those exist. I don't think there are any specific scoring systems for mental health issues or formalized assessments systems." Interviewer "How comfortable are physicians saying stay/go for young people with mental health issues?" "I would say uncomfortable. " Interviewer "How often do you think the physicians make the decisions themselves?" "I would say often enough because we see such a high volume of substance abuse and addictions or mental health excuse me that we can't refer everyone to psychiatry so we sort of have a process where we definitely will refer who we think are higher risk and the younger the person the more likely we are to refer and the comfort level is lower with younger patients." ED Physician, Participant #24. --- That Magical Number Participants spoke about young people knowing that when they were screened or assessed as a patient in the ED that a certain score or cut-off point was needed to either be admitted or get referred to services. They explained that they would use this knowledge to downplay symptoms if they wanted to go home, or on the other hand identify symptoms they knew would get them help. One participant explained their experience with this: "People kind of knew that the ER was assessing you and so there became this idea that people, if you want help then you have to make yourself look as bad as possible and make yourself show as many symptoms as possible so that you would meet that magical number. I think whatever assessment people are gonna use the 5 or 10 points before you hit that magical number or even that magical range also suck... I know that every assessment has young people who don't think it's that hard to fill out and there are people who fucking hate it so it's kind of hard to pick one in particular but for me it's more just what we do before what we do with the people who don't meet the criteria and how we can give them better resources 'cause they're clearly also not doing well, no one wants to waste their evening in a hospital." --- IYS staff, Participant #4. Physicians in the ED corroborated this, explaining that assessments and screening are used to determine if a young person is in crisis and the level of severity to determine if they should see a mental health specialist (such as a psychiatrist) for an evaluation. A physician explained this process from their perspective: --- "I'd say at least two-thirds of the kids we see actually do not require an emergent psychiatric assessment and in fact just needs some directions as to what the situation is, get some perspective on where, what kind of resources exist for their needs and then I talk to them about the various resources and the way I talk about three tiers, like the highest tier is the most emergent they need to see a psychiatrist really for the perspective of hospitalizing them, to keep them safe and stabilize until we can connect them to community resources. The second tier is when I will put in place a crisis response team and crisis type management whereby a psychiatrist or a mental health team provider will contact them the next day or so to see what the situation is and help them navigate the healthcare system and the mental health resources. " ED Physician, Participant #9. --- The Many Aspects of Health Participants identified specific assessments used in the ED (see Table 4 for further details) and acknowledged that ED assessments were inconsistent, especially regarding substance use. They encouraged recognizing factors beyond self-harm risk and incorporating a holistic approach to consider mental, physical, and sexual health, substance use, social determinants of health (such as housing and social supports) and culture. A parent shared her story of her experience with her child in the ED with regards to assessments, as well as recommendations for improvement: --- "I keep coming back to a substance use assessment done by a trained and experienced professional... A medical professional suggesting that a parent take a youth home who is high, drunk, suffering from suicidal ideation, taking prescription medication, soon to be coming down off their high -this is highly dangerous in my mind and experience. If the medical professionals and the parent have no clear idea of what they are trying to support or possibly will need to manage than how can anything other than having the youth stay in hospital where potential necessary treatment is immediately available even be considered, let alone suggested... A substance use assessment may include questions and observations related to identifying where challenges are existing and how problematic these challenges are. These areas may include but are not limited to family life, social experiences, school/ work life, mental health state, changes in behaviour, appearance, coordination (walking, sitting, writing), acute and chronic memory issues, red or glazed eyes, dilated or constricted pupils, unusual body odors, tremors, under-weight, low energy, bruising, dehydrated, dizzy, sleep related issues, etc. " Parent, Participant #26. --- Theme 3 Inadequate Communication Participants emphasized young people having to repeat painful or traumatic histories to multiple care providers and having providers deliver inconsistent or incorrect information when referring to other services. All communication between a young person and another person was considered at the interpersonal level of the SEM. In addition, communication between various services (different charting systems and medical records) made it a challenge for continuity of care for young people. The inconsistency and inability of charting systems to communicate across the province was considered a policy level factor. --- Person to Person The direct communication between health care providers and young people had a significant impact on a young person seeking help. A recurring concept discussed was that when young people present to an ED with a MHSU concern, they are in a very vulnerable state where they want or need help. They are then asked to repeat their story numerous times to each care provider throughout this process, which can be a difficult and traumatic experience. A participant expressed their frustration with young people having to repeat stories: In addition to repeating stories to different people, participants cited negative experiences when young people were not treated as a person first during conversations with care providers. This left youth in the ED feeling invalidated, unsafe, and stigmatized during an already traumatic experience. Participants acknowledged the importance of the communication with health care providers, as one negative experience can keep a young person from seeking out help again in the future. A participant shared their experience: "I went in [to the ED] and I was sharing the bit, struggling with eating disorder and I had someone respond "so you just want to be skinny, that's your issue"?... it just felt icky to me or I've had times where my lab work was very off and I had a nurse say to me "don't you realize you're killing yourself here, like this is very serious, do you not see how serious this is"? And again it wasn't in a compassionate light, I've had friends who had struggled more with I think like bulimia and when they opened up about their binging/purging there was this kind of horrified look on the physician's face and they shared with me how they just felt so embarrassed that they even admitted to what they were doing. I think sometimes we hope as people who've struggled when we share these things be met with this empathy and understanding." "If --- Peer Support Worker, Participant #4. A parent also shared their experience with a child in the ED and the "lack of trauma-informed care". She summarized her experience and compared the difficulties of mental health compared to physical health: "So, in general the experience of the process of acquiring support through the ED is quite stressful, unsupportive, difficult to navigate and not all validating. It is not as simple as going in and saying you need medical assistance and then being supportively brought through the appropriate steps of care, the way which you would if you went in for a broken limb or a high fever with severe nausea and emesis. " Parent, Participant #26. Participants also spoke about miscommunication regarding referrals from the ED to community IYS based on personal conversations of what they were told in the ED. This then led to unrealistic expectations upon referral with regard to wait times for services and what services are offered. A participant discussed the miscommunication about referral to other services: "So I think for me it's less the assessment and more like really encouraging people to do to like give those like good resources and not be lazy about it, I think there sometimes people who will give out a resource but they don't know anything about how to access that resource, they don't know if that resource has space the number of young people. I know [people] who've gotten a resource from the ER after being told they weren't going to get a bed for that night and then they call that place and there's a 3-year waiting list. " IYS staff, Participant #4. Participants also spoke about receiving pamphlets as a referral with no explanation beyond this which was not helpful communication. A parent summarized their experience: "We just get pamphlets and the pamphlets aren't helpful... For my son, when he's discharged, he's handed from PED [pediatrics] he's handed a bunch of pamphlets maybe a packet of pills that he just gobbles down, he's feeling rather hopeless, helpless, he's given a bunch of numbers to call, if that, which he doesn't, he's not able to he's just not able to, he had gone there for help but yeah wasn't able to get any. " Parent, Participant #17. --- Service to Service Participants identified communication when transitioning between services to be suboptimal, as there is a lack of communication between charting systems in different agencies (even within the same province). This again led to repetition of young people's stories (as discussed above in 'Person to Person' communication). Participants acknowledged the need for continuity of care through improved charting systems and referrals. A physician shared this from their perspective: "It's just a hope you say 'go to Foundry!' But no part of their chart, nothing follows them to there so they can go to the Foundry and Foundry may have no idea that they've even been to the emergency department or what their concerns were at the time or why they're even showing up. " ED Physician, Participant #1. Another health care provider discussed the difficulty with systems communicating and highlighted this, "continuative care just gets totally dropped, the only way it survives is if we're all on the same medical record so that's something that kind of keeps that continuative care alive" (Social Worker, Participant #5). --- Discussion Through our study and thematic analysis, we identified important factors at all levels of the Social Ecological Model (SEM) that need to be considered for young people with MHSU concerns presenting to the ED and referral to other community-based services such as IYS. While factors were identified at the individual level, these results could not be presented since no participants were considered the age of a young person (12-24 years) at the time of the study. At the interpersonal level, inadequate communication between ED staff and young people affected overall care and contributed to negative experiences when seeking help. Other studies in Canada have found that individuals of all ages presenting to the ED with MHSU concerns reported feeling stigmatized and discriminated against by staff in the hospital based on their communications with providers [35,36]. Hawk and colleagues acknowledged the importance of interpersonal relationships and the need for sympathetic health care staff in the ED after an overdose [24]. A systematic scoping review identified therapeutic alliance and health care staff showing empathy and non-judgement as being of high importance for individuals accessing substance use disorder treatment [37]. In Canada, recovery-oriented guidelines exist for mental health practice that recommend use of similar principles such as recognizing a person first, affirming autonomy, focusing on strengths, and building collaborative relationships [38]. However, Canadian emergency physicians have very limited formal MHSU training, often less than a few weeks in a single setting. Our results indicate that training for ED clinicians on MHSU, recovery principles, and traumainformed care could improve the care for young people. Purkey and colleagues have also acknowledged the need for training on trauma-informed care within the ED, as well as the need for institutional buy-in and have outlined detailed recommendations to do so [35]. At the organizational level, context of the environment was identified as important, with considerations for assessments, the ED and the hospital (specifically wait times, staffing issues, and the physical space). Regarding assessments, participants expressed a need for them to be youth-friendly, to be appropriate for MHSU concerns, and to consider health beyond symptoms and scoring. Within the ED in Canada, there is little guidance and great variation with respect to how to assess and treat young people presenting with MHSU concerns [30]. Two mental health assessment tools for young people have been developed in Canada: HEARTSMAP [31] and HEADS-ED [39]. Tools acknowledged for use of substance use assessment are the CRAFFT tool, which considers a variety of substances; AUDIT/AUDIT-C, which considers alcohol use disorders; and two questions on the NIAA tool [40,41]. However, participants acknowledged that youth-oriented assessments that do exist for MHSU concerns are not mandated or regulated in Canada; thus, use of them varies depending on the clinician. No guidelines or consensus statements from national organizations such as Canadian Association of Emergency Physicians or National Emergency Nurses Association exist. Newton and colleagues have also acknowledged that specialized instruments to screen for and diagnose MHSU problems for young people are not yet standard components of clinical assessments in the ED [42]. While various assessments exist for different reasons, our results indicate that the development of a comprehensive biopsychosocial assessment that is youth-friendly and considers mental health concerns, substance use, and other important social determinants of health such as housing and social supports would be highly beneficial for both ED clinicians and young people. Such an assessment would also need to be standardized for physicians working with this population in that it should be used consistently across EDs and IYS across the province (with findings shared with consent and referral) to enhance continuity of care and transitions between services. Wait times and the experience while waiting was a concern identified in this study. This could be due to limited resources such as staffing shortages and other care priorities, especially during the global COVID-19 pandemic [43]. It is likely that at the time of the study priority in hospitals was put on those presenting with COVID-19. However, most of our participants who spoke of lived ED experience referenced this from the past and likely prior to the pandemic. Participants spoke of the benefit of having psychiatric nurses in the ED setting to potentially address some of these concerns. Work has been done within Canada to better understand this role, and patients, families as well as ED staff in general valued such an addition in the ED [44]. However, such specialized staff vary in Canada and are only embedded within certain hospitals. Thus, without adequate resources (such as specialized MHSU staff including psychiatric nurses) in the ED, it warrants consideration of ED staff capacity to do thorough assessment, treatment, and referral in a youth and person-centered way. One of our participants spoke of their personal lived experience and acknowledged seeing burned-out staff in the ED. Since EDs may be the only access point for care of young people with MHSU concerns after IYS hours of operation and in communities without IYS or MHSU services, hospitals and staff need to be better equipped to deal with young people presenting with concerns. One consideration is the implementation of mental health staff within EDs. This could include peer support workers who may utilize their own lived experience to aid in people presenting with concerns. A study has been done to explore the potential in Australia acknowledging the important skills peer support workers could bring [45]. Also in Australia, a new model of mental health liaison nursing care has been implemented and evaluated within EDs [46][47][48]. Some of the key guiding principles of this model include a lead nurse practitioner/consultant to a designated team of specialist mental health nurses during extended hours 7 days per week to work closely with other ED staff and complement psychiatry services [46]. Initial evaluation of this is promising, showing a reduction in ED length of stay, less wait times to see a professional, referrals to community mental health teams, and clinician and patient acceptability [46]. Patients acknowledged the therapeutic benefits including being listened to and understood [47,48]. Implementation of such a model in Canada would require significant resources and organizational change. However, as our results indicate, the EDs are currently significantly under resourced, especially for people presenting with MHSU concerns. The authors conducting this work in Australia acknowledge the need for resources, structural considerations within the ED, and connections to other mental health services [47] which would all be necessary considerations for a similar model to be used within Canada. With regard to the community

Background: Integrated youth services (IYS) provide multidisciplinary care (including mental, physical, and social) prioritizing the needs of young people and their families. Despite a significant rise in emergency department (ED) visits by young Canadians with mental health and substance use (MHSU) concerns over the last decade, there remains a profound disconnect between EDs and MHSU integrated youth services. The first objective of this study was to better understand the assessment, treatment, and referral of young people (ages 12-24 years) presenting to the ED with MHSU concerns. The second objective was to explore how to improve the transition from the ED to IYS for young people with MHSU concerns.We conducted semi-structured one-on-one video and phone interviews with stakeholders in British Columbia, Canada in the summer of 2020. Snowball sampling was utilized, and participants (n = 26) were reached, including ED physicians (n = 6), social workers (n = 4), nurses (n = 2), an occupational therapist (n = 1); a counselor (n = 1); staff/leadership in IYS organizations (n = 4); mental health/family workers (n = 3); peer support workers (n = 2), and parents (n = 3). A thematic analysis (TA) was conducted using a deductive and inductive approach conceptually guided by the Social Ecological Model.We identified three overarching themes, and factors to consider at all levels of the Social Ecological Model. At the interpersonal level inadequate communication between ED staff and young people affected overall care and contributed to negative experiences. At the organizational level, we identified considerations for assessments and the ED and the hospital (wait times, staffing issues, and the physical space). At the community level, the environment of IYS and other community services were important including wait times and hours of operation. Policy level factors identified include inadequate communication between services (e.g., different charting systems and documentation).

peer support workers could bring [45]. Also in Australia, a new model of mental health liaison nursing care has been implemented and evaluated within EDs [46][47][48]. Some of the key guiding principles of this model include a lead nurse practitioner/consultant to a designated team of specialist mental health nurses during extended hours 7 days per week to work closely with other ED staff and complement psychiatry services [46]. Initial evaluation of this is promising, showing a reduction in ED length of stay, less wait times to see a professional, referrals to community mental health teams, and clinician and patient acceptability [46]. Patients acknowledged the therapeutic benefits including being listened to and understood [47,48]. Implementation of such a model in Canada would require significant resources and organizational change. However, as our results indicate, the EDs are currently significantly under resourced, especially for people presenting with MHSU concerns. The authors conducting this work in Australia acknowledge the need for resources, structural considerations within the ED, and connections to other mental health services [47] which would all be necessary considerations for a similar model to be used within Canada. With regard to the community level, the context of the environment of IYS and other community services were identified as important. Specifically, this included wait times and hours of operation. Important policy level factors identified include inadequate communication between services (e.g., different charting systems and documentation) and context of the environment when considering the ED and hospital-specifically shortages of detox beds and specialized units for young people with MHSU concerns. Systemic changes to address these concerns include a universal charting system within the province, more detox facilities or designated detox spaces embedded within the hospital, and perhaps specialized MHSU units for young people. --- Future directions The results indicate there are systemic and long-term policy changes needed including development of a MHSU assessment for young people, specialized units, detox beds, increased staffing, and some aspects of the physical environment (e.g., paint, lighting, chairs). However, the results also indicate that there are certain areas of improvement to be addressed immediately. This includes communication between health care providers and young people, the experiences of young people waiting in the ED, and potential implementation of a combination of existing mental health, substance use, and suicide risk assessments. The working team (authors) for this project (ED2Foundry) has received funding to implement a small-scale intervention in two EDs in BC to enhance the experience of young people with MHSU concerns and referral to IYS in the community such as Foundry. The team is currently using the results from this study to consult with young people and ED staff to develop, implement, and evaluate an intervention. --- Strengths and limitations The study had a wide range of stakeholders with lived experience; however, not having the firsthand perspective of young people at the time of the study was a limitation. The authors attempted to enroll young people as participants and asked all participants if they could share study details with their networks. The study team was contacted by one young person who participated in an interview; however, the interview was inaudible, and the data could not be used. This limited contact could be perhaps due to this group having increased MHSU concerns actively at the time of the study and not being able to participate. This study occurred in the early stages of the COVID-19 global pandemic. In attempt to reconcile this, the team worked to obtain the perspectives of young people through parents, who shared both their own experiences and their child's experiences, and current adults shared their own MHSU experiences from when they were young. For health care providers (other than the ED physicians and IYS staff ) not all disclosed whether they worked in the ED or the community when demographic data was collected thus the authors could not tease this apart when interpreting results. The study was guided by a theoretical framework [21]. Several steps were taken by the authors described in the data analysis section to ensure rigour [28]. The authors also consulted the 15-point "checklist" for good thematic analysis [26]. The authors are currently discussing the results with youth with lived/living MHSU experience to determine the future intervention aspect. --- Conclusion This study identifies important factors to improve the ED care experience of young people presenting with MHSU concerns and referral to appropriate services such as IYS. This includes long-term systemic changes such as the development of a MHSU assessment for young people and aspects of the hospital and ED environment such as increased staffing, specialized units and staffing and the physical space. The results also indicate areas of improvement to be addressed immediately, such as communication between health care providers and young people, and the current wait time experiences of young people while in the ED when resources are limited. This study supports intervention in the ED for young people presenting with MHSU concerns and the authors are using the results of this study to inform intervention development and implementation. --- Availability of data and materials The datasets generated and analyzed during the current study are summarized in this manuscript. The raw data are not publicly available due to the possibility of confidentiality of participants being compromised. Contact the corresponding author krista. glowa cki@ ubc. ca should you have any questions about access to data. --- Abbreviations IYS: Integrated youth services; ED: Emergency department; MHSU: Mental health and substance use; BC: British Columbia; SEM: Social Ecological Model. --- Authors' contributions FS, SB, DB, SM designed the study and attained funding. MW, JW, SB, and FS collected data. KG and MW completed data analysis and completed writing of the manuscript. KM, FS, SB, and DB contributed to writing of the manuscript. All authors read and approved the final manuscript. --- Declarations Ethics approval and consent to participate The study protocol was approved by the University of British Columbia Ethics Board (H20-01396), in Vancouver, British Columbia, Canada. The study was performed according to the standards of the Declaration of Helsinki (1964) and its later amendments. Informed verbal consent was obtained from participants prior to study participation. --- Consent for publication Informed verbal consent was obtained from participants for publication. --- Competing interests Not Applicable. • fast, convenient online submission • thorough peer review by experienced researchers in your field --- • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year --- • At BMC, research is always in progress. --- Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research? Choose BMC and benefit from:? Choose BMC and benefit from: --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: Integrated youth services (IYS) provide multidisciplinary care (including mental, physical, and social) prioritizing the needs of young people and their families. Despite a significant rise in emergency department (ED) visits by young Canadians with mental health and substance use (MHSU) concerns over the last decade, there remains a profound disconnect between EDs and MHSU integrated youth services. The first objective of this study was to better understand the assessment, treatment, and referral of young people (ages 12-24 years) presenting to the ED with MHSU concerns. The second objective was to explore how to improve the transition from the ED to IYS for young people with MHSU concerns.We conducted semi-structured one-on-one video and phone interviews with stakeholders in British Columbia, Canada in the summer of 2020. Snowball sampling was utilized, and participants (n = 26) were reached, including ED physicians (n = 6), social workers (n = 4), nurses (n = 2), an occupational therapist (n = 1); a counselor (n = 1); staff/leadership in IYS organizations (n = 4); mental health/family workers (n = 3); peer support workers (n = 2), and parents (n = 3). A thematic analysis (TA) was conducted using a deductive and inductive approach conceptually guided by the Social Ecological Model.We identified three overarching themes, and factors to consider at all levels of the Social Ecological Model. At the interpersonal level inadequate communication between ED staff and young people affected overall care and contributed to negative experiences. At the organizational level, we identified considerations for assessments and the ED and the hospital (wait times, staffing issues, and the physical space). At the community level, the environment of IYS and other community services were important including wait times and hours of operation. Policy level factors identified include inadequate communication between services (e.g., different charting systems and documentation).

prevalence by 2045 of all global regions [1]. In highincome countries, socioeconomically disadvantaged communities have been disproportionally affected [2]. The beneficial role of physical activity (PA) in the prevention of T2D has been well established [3] and more people engaging in regular PA could help in curbing this growing pandemic. For low-income sub-Saharan African countries, implementation of PA is a feasible prevention strategy. Incorporating regular PA in people's daily lives, however, remains a challenge. Sustainable behavior change is known to be complicated and dependent on many factors including a supportive physical and social environment [4]. Behavioral theories have been shown useful in the adoption of regular PA, but most evidence to support such theories has originated from populations that are Western, educated, industrialized, rich and democratic (WEIRD), and as such, have been called "frequent outliers" -i.e., the least representative populations in terms of human psychology and behavior [5]. Cross-sociocultural validation of behavioral theories among non-WEIRD populations is therefore urgently needed. Self-Determination Theory (SDT) has offered promising insights in people's engagement in sustainable lifestyle behavior such as performing regular PA [6]. SDT argues that the quality of people's social environment plays a crucial role in maintaining such behavior. SDT distinguishes between autonomous forms of motivation which emanate from within oneself or from abiding values and controlled forms of motivation which are triggered by sources external to the actual behavior [7]. Social contexts satisfying an individual's perceived competence, autonomy, and relatedness (defined as the basic psychological needs) have been shown to foster more autonomous types of motivation resulting in a more sustained behavior change [7]. SDT has been claimed to be etic universal, meaning that its cross-cultural validity can be empirically identified [7]. However, existing studies on cross-cultural generalizability have been focused on comparing industrialized countries in high-and middle-income regions [8]. And, to our knowledge, no studies comparing SDT's generalizability between sub-Saharan Africa and other settings have been published yet. Another limitation is that most of the existing studies on cross-cultural generalizability have focused on specific domains such as education, work and well-being [9][10][11]. While evidence supports SDT with regards to engagement in sustained PA [6], cross-cultural validation has been scarce and limited to students or athletes [12][13][14]. A similar conclusion can be drawn for social support and self-efficacy, two constructs that have been consistently linked to PA, including in low-and middle income settings [15,16]. Evidence for these associations, however, typically comes from single country analyses which may not guarantee cross-cultural generalizability. Moreover, studies have typically zoomed in on one of both factors, ignoring the possible interactions between concepts. To address this last shortcoming, we recently assessed an SDT-based process model with integration of social support and self-efficacy and with PA as the intended behavior in a rural Ugandan population [17]. The study showed a positive relationship between the frequency of vigorous PA and identified regulation, a form of autonomous motivation elicited through associating PA with an individuals' goals or values, such as "being healthy". Further in line with SDT, identified regulation operated as a mediator between vigorous PA and barrier self-efficacy and perceived social support, which show conceptual and statistical parallels with, respectively, perceived competence (i.e. one's sense of efficacy with respect to both internal and external environments) and perceived relatedness (i.e. the sense of being supported by significant others in one's actions) which we defined earlier as the psychological needs in SDT [7,[17][18][19]. While this study provided evidence on SDT and PA in sub-Saharan Africa, cross-cultural validation of the theory requires explicit comparison between different settings. The objective of this study is to compare an adapted version of this process model across socio-economically disadvantaged populations in two sub-Saharan African and one European country, using state-of-theart modeling techniques. First, we will test if it is feasible to compare motivational determinants of PA (i.e. autonomous motivation, self-efficacy and social support) across diverse socio-cultural environments. Second, we will compare mean levels of these determinants across the three settings. Third, we will test the hypothesized motivational process model in each of the three settings and across the settings. Finally, we will investigate the relationship between PA and HbA1C in a separate model with the aim to connect the self-reported PA outcomes with a more objective measure. The hypothesized model for this study (see Fig. 1), assumes: (1) a positive association between identified regulation (i.e. a form of autonomous motivation) and PA; (2) a positive association between perceived social support, self-efficacy and identified regulation; (3) a positive total effect of perceived relatedness and self-efficacy on PA outcomes (which includes the associations in the previous steps); and (4) a positive association between PA and HbA1c. --- Methods --- Study design and procedures This study contributes to the validation of a theorydriven framework that guided the implementation of the SMART2D trial (an adaptive implementation trial to improve self-management and to promote a healthy lifestyle among people at risk of or living with T2D in Uganda and South Africa and a feasibility implementation trial in Sweden) [4]. The study used cross-sectional baseline data collected from two rural districts in eastern Uganda, a peri-urban township in the Western Cape in South Africa, and two socio-economically disadvantaged districts of Stockholm in Sweden. --- Study settings The rural Ugandan population was characterized by a collectivist society with low levels of migration and a high proportion working in agriculture. The urban South African population was characterized by national and international migrant workers with a relatively high unemployment rate. Participants of this setting reported that frequent migration hindered them to build strong community ties. The urban Swedish population consisted of a high proportion of immigrants (approx. 60%) with a diversity in culture and ethnic background living in a society where health and lifestyle are individualized. All three populations were socio-economically disadvantaged in several aspects, but with a sharper socio-economic inequality in the South African setting. More details about the social and built environment, the health system and the population of the study sites can be found elsewhere [4]. --- Study participants, sampling and recruitment Study participants were considered eligible if they had resided in one of the study sites for at least 6 months; were aged 30-75 years; had not been previously diagnosed with T2D for longer than 12 months (for the Ugandan and South African site) or 5 years (for the Swedish site); and had a confirmation of prediabetes or diabetes. Pregnancy and serious mental disability were exclusion criteria. In Uganda, 712 participants were recruited by trained field research assistants approaching households in the study area in a random manner. In South Africa, 566 participants were recruited from two community health centers located in the township upon referral by a health care worker. In Sweden, 147 participants were recruited through screening in public spaces and facilitybased screening in two primary health centers. Consenting participants were screened through a fasting plasma glucose test in Uganda, a random plasma glucose test in South Africa and the Finnish Diabetes Risk Score (FIND-RISC) in Sweden, except for diabetes patients recruited directly from the health facility in the Swedish setting. Confirmation of T2D or pre-T2D was done through fasting plasma glucose tests in Uganda and South Africa (at least two tests <unk>6.1 mmol/L for pre-T2D and at least two tests <unk>6.9 mmol/L for T2D) and through an HbA1c test in Sweden (HbA1c <unk> 42 mmol/mol for pre-T2D and HbA1c <unk> 48 mmol/mol for T2D). More details about the selection criteria and the recruitment process can be found elsewhere for the Ugandan setting [17], the South African setting [20] and the Swedish setting [21]. --- Data collection A questionnaire was administered by trained field workers and included socio-demographic items, PA-and motivation-related scales, anthropometric and biochemical measurements. Data were collected between January 2017 through December 2017 in Uganda, between August 2017 and November 2018 in South Africa and between June 2017 and January 2019 in Sweden. --- Measures Identified regulation towards physical exercise was assessed through the Treatment Self-Regulation Questionnaire for people with diabetes. This scale has been widely used to test PA self-regulation and studies have reported adequate reliability [22]. Guided by factor loadings identified in the study by Levesque et al. [22], four items were selected to measure identified regulation (see Additional file 1). Participants responded to each item on a 5-point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). To measure perceived social support, an adapted version of the scale for participation and involvement of family members and friends in PA was used [23]. This scale has been used and validated in a variety of contexts [23]. Five items of the initial measure were selected based on their presumed cross-cultural adaptability and factor loadings in previous studies (see Additional file 1). Participants responded to each item on a 4-point Likert-type scale ranging from 1 (never) to 4 (more than once a week). Perceived social support shows conceptual parallels with perceived relatedness and the same scale was used by others to measure perceived relatedness [24]. To emphasize the concept of perceived support among the study participants, we introduced the questions with the following statement: --- "We want to understand to what extent people close to you (friends, family or relatives) have helped you to do physical activity". Barrier self-efficacy (or self-regulatory efficacy) corresponds to the perceived capability to maintain PA given various conditions or impediments (i.e. barriers). Six items were adapted from the health-specific self-efficacy scale developed by Schwarzer et al. (2007) (see Additional file 1). Barriers included in the original questionnaire were modified to barriers relevant to the study contexts. Participants responded to each item on a 5-point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). Self-efficacy and perceived competence have shown to be correlated [25], but a conceptual difference needs to be acknowledged [18]. Unlike self-efficacy, perceived competence encompasses the concept of personal effectance, or the perceived need to effect change and attain valued outcomes [18]. PA was measured through: 1) self-reported frequency of vigorous PA; and 2) self-reported frequency of moderate PA. Initial questions from the World Health Organisation "STEPS" survey [26] were contextually adapted (see Additional file 1). Different measures were chosen since associations with SDT constructs may depend on the intensity of PA [6]. More detail on the motivational and PA measures can be found in a previous study on SDT by De Man et al. [17]. HbA1c was measured using capillary blood samples obtained with a Point-of-care HbA1c Analyzer Cobas b101 (Roche Diagnostics) with the respective test and control reagents. Sociodemographic characteristics were self-reported based on the categories included in Table 1. Height and weight were measured using standard procedures. --- Contextual adaption All measures were translated into the local language of the study populations (i.e. Lusoga, Swedish, Arabic, Somali and isiXhosa), and adapted to the context based on inputs from a team of local research assistants. Measures were then back translated to English and adjustments made where necessary to ensure that the meaning of the questions was not lost. Local validity was ensured through piloting in a non-study area, training of data collectors (e.g. through mock interviews), and minimizing inter-interviewer variability. --- Data analysis The current state-of-the-art approach to compare mean levels and associations of latent constructs across different settings is multigroup structural equation modelling (MGSEM) [27]. A major condition to apply this technique is measurement invariance of constructs across different settings [27]. Measurement invariance supports the idea that subjects of different subgroups have a similar understanding and give a similar meaning to the items of a latent construct. Testing for measurement invariance was based on subsequent steps imposing additional constraints to the models. Before testing for measurement invariance, separate measurement models were assessed for each construct and country separately. Confirmatory factor analysis (CFA) was used to assess the loadings of the item indicators on the latent variables (i.e. identified regulation, barrier self-efficacy, and social support) and the goodness of fit of these measurement models. Subsequently, simultaneous analysis of equal form (i.e. configural invariance), equivalence of factor loadings (i.e. metric invariance) and equivalence of intercepts (i.e. scalar invariance) was conducted across countries through MGSEM. In case these measurement models would yield an acceptable fit and were shown invariant, the fit of the hypothesized structural equation model (SEM) was assessed per country separately. Finally, to test if the associations between the constructs across the three countries were similar, we compared the difference in model fit between a model constraining these associations as similar across settings and a model without constraining a specific association. In case model fit was not significantly worse between these nested models, we concluded that that specific association was similar across settings. Model fit was evaluated based on multiple indices, including root-mean-square error of approximation (RMSEA) corrected for nonnormality [28], with target values as proposed by Hu and Bentler [29]: the comparative fit index (CFI) <unk> 0.95, Tucker-Lewis index (TLI) <unk> 0.95, the root mean square error of approximation (RMSEA) <unk>0.06, and the standardized root mean square residual (SRMR) <unk>0.08. Criteria indicating a significant decrease in model fit for the more constrained model (i.e. indicating non-invariance) were as follows: a decrease in CFI larger than 0.01 combined with an increase in RMSEA smaller than 0.015 or a non-significant scaled <unk>-square difference test [30]. Since items' distributions departed from normality, we used maximum likelihood estimation with robust standard errors and a Satorra-Bentler scaled test statistic [31]. Covariates were added to address potential sources of confounding based on theory and identified through directed acyclic graphs. Education, marital status and body mass index (BMI) were included to adjust the motivational constructs. Age, sex, occupation, BMI and education were included to adjust the PA outcomes [15]. Data were analyzed using R software with the packages "lavaan" and "semTools". To study the link between PA outcomes and HbA1c, a separate linear regression model was used controlling for the following covariates: age, sex, BMI and reported intake of oral antidiabetic medication [32]. --- Missing data Missing data for the Ugandan site varied from 0.0-1.3% per variable, for the South African site from 0.0-1.4% and 2.0-10.2% for the Swedish site. The variable responding to the question: "are you currently on any oral hypoglycemic agents?" was missing more frequently: 49.4, 51.2 and 14.3% among participants in the Uganda, South African and Swedish site respectively. BMI was missing among 16.3% of the participants in the Swedish site. Multivariate imputation by chained equations with predictive mean matching was used to handle the missing data under a missing at random assumption. Rubin's rules were used to pool point and SE estimates across 30 imputed data sets. The procedure was done using the "Mice" package in R. For the variable regarding oral hypoglycemic treatment, a sensitivity analysis was ran ignoring this variable. --- Table 1 Demographics, diabetes parameters and physical activity behavior of the study population per setting Legend: HbA1c Hemoglobin A1c, PA Physical activity, BMI Body Mass Index. a reported in [17]. b reported in [20], except for HbA1c and PA. c reported in [21], except for HbA1c and PA --- Uganda a South Africa b Sweden c --- Mean (SD) Age in years 52.5 ( --- Ethics approval The study was approved by the ethics committees in each of the respective countries (See declarations). Informed consent was obtained from all individual participants included in the study. --- Results --- Description of the study samples Essential characteristics of the samples are summarized in Table --- Measurement models and invariance As mentioned above, we first intended to obtain adequate measurement models per country setting based on the same items. The proposed measures for social support and barrier self-efficacy did not yield an acceptable model fit across the three settings (see Additional file 1). To obtain acceptable measurement models, we considered the models' modification indices and items' factor loadings. This resulted in. exclusion of item 2 for the social support construct (see Additional file 1), exclusion of items 1 & 2 for the barrier self-efficacy construct, and adding correlated errors between item 3 & 4 for identified regulation (the items of this construct were mixed up with other items to minimize acquiescence bias, but items 3 & 4 followed after each other). Despite piloting the questionnaire, some participants reported item 1 of the barrier selfefficacy construct as confusing. The two other excluded items may have been interpreted as more directive than intended in the South African setting (i.e. "being told" instead of "encouraged"). Model fit of the resulting measurement modelswas good to excellent for the three settings (see Table 2). Unstandardized factor loadings were all significant (z > 1.96) and standardized factor loadings were higher than.50 (see Additional file 1), except for barrier self-efficacy. For this construct, two items had very low factor loadings (item 6 <unk> =.23; P =.115 and item 5 <unk> =.32; P =.039) in the Swedish setting. We ran a sensitivity analysis with a construct without item 6 which did not result in major differences in the results to follow. In a second step, multigroup confirmatory factor analysis was conducted across the three settings assessing the difference in fit for different levels of invariance. The three constructs showed configural invariance and metric invariance across settings (see Table 2). Social support and identified regulation showed partial scalar invariance (for each, 1 response item was excluded from constraining equal intercepts). The barrier self-efficacy construct showed full scalar invariance. --- Comparison of latent means of motivational constructs The level of invariance established in the previous section (i.e. partial to full scalar invariance) allowed us to compare latent mean estimates across settings. Latent mean estimates were highest in Uganda and lowest in South Africa (see Table 3). The difference in identified regulation between South Africa and Sweden was small. Compared to the other settings, social support was substantially higher in Uganda. Compared to the other settings, barrier self-efficacy was substantially lower in South Africa. --- Structural models The hypothesized structural model was fitted for each country separately. This resulted in excellent model fit in all three settings (Table 4). However, associations differed substantially across settings. Identified regulation was only associated with vigorous PA in Uganda and with moderate PA in South Africa. Social support was associated with PA outcomes in all three settings. Barrier self-efficacy showed to be associated with identified regulation in Uganda, but not in South Africa. In Sweden, we found a positive association which was not significant, possibly due to a relatively small sample size (standard errors in Sweden were 0.252; vs. Uganda 0.035). Social support was associated with both PA outcomes in Uganda, with vigorous PA in South Africa and with moderate PA in Sweden. Barrier self-efficacy was associated with moderate PA, but only in South Africa. In Sweden, we found a positive association which was not significant, again, possibly due to a relatively small sample size (N = 147). Metric invariance was established for all constructs across the three settings which supports a meaningful comparison of associations between these constructs across settings. The same structural model as for the single country settings was used to build a multi-group structural model. Based on the <unk>2-difference test, this model did result in a significantly worse fit compared to a model in which one of the associations mentioned in Table 4 (e.g. identified regulation and PA) was freely estimated across the three settings, except for the association between social support and identified regulation (for vigorous PA: <unk> 2 scaled = 2.81; df = 2; p-value =.245 and for moderate PA: <unk> 2 scaled = 2.76; df = 2; p-value =.375). In other words, constraining the latter association as equal across countries did not result in a significantly worse fit, indicating a similar association. The other associations, however, showed to differ across settings. --- Relationship between HbA1c and types of PA In all three settings, self-reported frequency of vigorous PA (i.e. number of days a week) was negatively associated with HbA1c scores (see Table 5). A regression estimate of -0.103 corresponds to the reduction in participants' HbA1c per extra day of self-reported vigorous PA per week. Associations with self-reported frequency of moderate PA were negative, but not significant in all three countries. As mentioned above, the variable regarding --- Discussion The aim of this study was to compare identified regulation, self-efficacy and social support and their association with PA and HbA1c across divers socio-cultural environments. Comparison of these motivational constructs was possible since measurement invariance could be established across samples, suggesting a similar understanding across study populations. However, obtaining adequate measurement models required modification of the initial measures (e.g., items had to be dropped). Estimates of the motivational constructs were highest in Uganda and lowest in South Africa, with a substantial difference for barrier self-efficacy and social support. Structural models did not correspond across settings. Identified regulation was positively associated with vigorous PA in Uganda and with moderate PA in South Africa. In Sweden, none of the PA outcomes was associated with identified regulation. The strength of the association between social support and identified regulation was similar across settings. Depending on the setting, the association between social support and PA outcomes was weak to insignificant. The association between barrier self-efficacy and PA outcomes was not significant. Self-reported PA was highest in Uganda and lowest in Sweden. Vigorous PA was associated with lower HbA1c across countries, while this association was not significant for moderate PA. The latent mean level of barrier self-efficacy was much lower in South Africa compared to both other settings, which may have occurred for several reasons. External barriers that are more prevalent in South Africa may hinder people from doing PA, including lack of security in the South African township setting [4]. South Africa's historical context and pervasive social inequality may also have contributed to the perceived difference in levels of self-efficacy [4]. The lower socioeconomic status of study participants, including many internal migrants, may have affected people's self-esteem [33], and consequently, their self-efficacy [34]. Potentially, this could also explain the lower self-efficacy level (compared to the Ugandan setting) in the Swedish setting, as the study sample was socioeconomically disadvantaged and included 60% migrants [35]. Social support was perceived to be much higher in the Ugandan setting compared to the other settings. This could be explained by the stronger social ties apparent in the rural Ugandan community versus the urban sample with many migrant workers in the South African and Swedish setting [4]. The proportion of participants that indicated to be married or co-habiting was also substantially higher in Uganda. In qualitative interviews conducted during the formative phase of the SMART2D project, participants from the Swedish setting reported to perform PA on their own, rather than with others, which corresponds to health and lifestyle being more individualized in Sweden [4]. We hope that future research can explore these factors in more depth. The positive association between identified regulation and PA (vigorous PA in Uganda and moderate PA in South Africa) is in line with previous studies on SDT and PA [6]. However, this association was not found in Sweden, nor for other PA outcomes. We assume this lack of association can be explained by the PA-related questions covering all types of PA, including traveland work-related PA. It is likely that most of the selfreported PA related to travel and work, especially in the African settings [36], which implies an important effect of triggers external to the activity. The difference between Uganda and South Africa could be explained by a different attitude towards the intensity types of PA, although this hypothesis was not tested in this study. In the rural Ugandan setting, participants may have been more used to performing vigorous activity as 69% of the participants were farmers [17], while in the urban township setting of South Africa, participants reported to perform less vigorous physical activity. As they seem less used to perform vigorous PA, they might perceive it as more demanding and hence, they might be more inclined to connect moderate rather than vigorous PA with autonomous forms of motivation. Other studies have highlighted the importance of the type of PA influencing the relationship between PA and the type of motivation [6]. For example, PA of a more repetitive nature was shown to be stronger associated with identified regulation than with intrinsic motivation [6]. Our study suggests that the association with more autonomous forms of motivation may depend on the perceived intensity of PA and that this association is context dependent. From a statistical perspective, differences in dispersion of the outcome data may explain why certain associations did or did not occur. For instance, vigorous PA in Sweden (inter-quartile range = 1.5) and in South Africa (interquartile range = 2) had a lower dispersion compared to moderate PA in both countries (inter-quartile range = 6 and 4 respectively). In Sweden, 69% of the participants reported to perform 0 days of vigorous PA. The construct of social support showed a similar positive association with identified regulation across the three country settings. Since this construct shows conceptual parallels with the concept of perceived relatedness [24], this finding supports the etic validity of the basic psychological needs theory, which posits that satisfaction of psychological needs fosters more autonomous forms of motivation and more sustainable behavioral outcomes (Deci and Ryan, 2000). However, this did not apply to the construct of barrier self-efficacy, which showed a positive relationship with identified regulation in Uganda, but not in South Africa and Sweden (although the latter could have been due to a lack of power). Moreover, barrier selfefficacy, which has been shown a consistent predictor of PA [16], did not show an association with PA outcomes. This may be explained by most of PA being related to work or travel, with an important influence by external triggers and not leaving much flexibility to participants. On the other hand, our data did provide support for an association, although small, between social support and self-reported PA. A potential explanation could be that participants included companionship at work or during travel in their conceptualization of social support. Higher self-reported PA in the African sites of the study compared to a high-income Western country such as Sweden is in line with global trends and likely due to a higher level of travel-and work-related PA [36]. The high levels of PA in Uganda is also in line with a recent national survey [37]. In sub-Saharan Africa, urban regions have been associated with a more sedentary lifestyle which may explain lower self-reported PA in the South African versus Ugandan setting [16]. Besides these global trends, a substantially higher proportion of the Ugandan participants reported to be employed. Self-reported frequency of vigorous PA showed a similar negative association with HbA1c across the three settings. Regression estimates for self-reported frequency of moderate PA were about half the size of vigorous PA estimates and non-significant. This is not in line with a recent meta-analysis of randomized trials which found changes in HbA1c driven by the duration of PA in a linear manner and independent of the type and intensity of the PA intervention [32]. If only duration and not intensity of PA would matter, a straightforward explanation of our findings would be that participants' reporting of PA is dependent on the intensity of PA. In other words, participants' perception to have performed 30 min. of moderate PA may be different from their perception to have performed 15 min. of vigorous PA. These findings from a real-life setting warrant further investigation as they would be crucial to consider in health promotion. Experimental trials using objective and self-reported measures can bring more insight. --- Study limitations and recommendations Comparison of motivational constructs was possible across settings, but measures had to be modified to obtain adequate measurement models. This incompatibility across very different contexts could be explained by translation to local languages altering certain nuances, a different contextual relevance of certain items and a different understanding by participants from different settings. As mentioned in the results section, two factor loadings of the barrier self-efficacy construct were very low for the Swedish site. However, sensitivity analysis after exclusion of the item with the lowest factor loading did not reveal any major differences. We further need to acknowledge that sampling procedures and criteria were different in the three settings for pragmatic reasons. As we documented in the methods section, participants in the Ugandan setting were recruited through a random sampling procedure, while participants in the other sites were mainly recruited from health centers, at the cost of reducing external validity. Moreover, selection criteria also differed across sites: participants from the Ugandan and South African site had not been diagnosed with T2D for longer than 12 months, while this was 5 years for the Swedish site. Such differences may have influenced the mean levels and relationships between the motivational constructs in this study. Another limitation of our cross-comparison study was that we did not directly assess relevant cultural differences (e.g. via measures designed to assess individualism/ collectivism) which could have been served as possible moderators of the effects we measured. " The distinction between self-reported vigorous and moderate PA offered an interesting perspective, but also hindered the association of constructs with other factors that determined PA performance. A distinction between categories of PA (e.g., work-, leisure-, transport-related PA) and other categories of motivation from the SDT continuum may add further insight into the role of motivation. In addition, controlling for other factors (e.g. perceived safety, availability of sports infrastructure, etc.) may result in a more nuanced image of the role of motivation. This study aimed to assess the cross-cultural validity of an SDT process model across different settings. However, the cross-sectional design of this study does not provide evidence for causal pathways or trends over time. While this study focused on people who were recently diagnosed, different dynamics may appear in people with long-standing diabetes. Studies collecting data at different time points and intervention trials can address these shortcomings. Finally, we need to acknowledge that the use of selfreported measures exposes our findings to bias, including social desirability bias, recall bias (people who value PA more as beneficial for health, may also have reported higher values of PA) and interviewer bias. Shared method variance between measures may have led to overestimation of associations. Objective registration of PA through a pedometer or accelerometer could have made our findings more robust and challenged self-reporting. --- Conclusion This is the first study comparing a motivational process model between Western and sub-Saharan African settings. Our findings suggest a similar understanding of these constructs across very different settings which makes comparison meaningful. Perceived social support and barrier self-efficacy levels showed to be lower in the urban samples with a high proportion of migrants in South Africa and Sweden, suggesting them to be psychosocially more vulnerable. Except for the association between perceived social support and identified regulation, the motivational process model was different across settings. Identified regulation showed stronger associations with PA outcomes than socials support and self-efficacy. However, we found that these relationships were dependent on the perceived intensity of PA, and hence, do not necessarily reject the etic validity of the underlying behavioral theory. We recommend more specific and objective PA outcomes to analyze these associations. Self-reported vigorous PA was related to lower Hba1C values across the three settings, while this was not the case for moderate PA. This discrepancy urges further exploration of people's perception of moderate and vigorous PA as it may have major implications for health promotion and education. Our study showed it is feasible to compare a sophisticated motivational model across very different settings. We encourage further research on the cross-sociocultural validation of such models. --- Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. --- Abbreviations BMI: Body mass index; CFA: Confirmatory factor analysis; CFI: Comparative fit index; FINDRISC: Finnish Diabetes Risk Score; Hba1C: Hemoglobin A1c; MGSEM: Multigroup structural equation modelling; PA: Physical activity; RMSEA: Root-mean-square error of approximation; SDT: Self-determination theory; SRMR: Standardized root mean square residual; T2D: Type 2 diabetes; TLI: Tucker-Lewis index; WEIRD: Western, educated, industrialized, rich and democratic. --- Supplementary Information The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s12889-022-12539-9. --- Additional file 1. Authors' contributions PA, MD, JV, TP, PD, JV, DG, TP, and JD played a major role in the conception of the study. PA, MD, TP, PD, DG, TP, and JV played a major role in the design of the study. FK, PD, KA, PD, DG, MD, and TP were in charge of the data acquisition. JD created the first draft of the manuscript and analysed and interpreted the data. FK, PD, KA, HA, PA, EW, MD, DG, JV, TP, HT, and RR substantially contributed to the interpretation of the data. FK, PD, KA, HA, PA, EW, MD, DG, TP, RR, HT, and JV critically revised the manuscript for important intellectual content, and read and approved the final version. --- Declarations Ethics approval and consent to participate "All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. " Informed consent was obtained from all individual participants included in the study. The study was approved by the ethics committees in each of the respective countries. In Uganda, this was the Higher Degrees, Research and Ethics Committee (HDREC) of Makerere University School of Public Health and the Uganda National Council for Science and Technology (Ref. HDREC-331 and HS 1917 respectively). In South Africa this was the University of the Western Cape Biomedical Science Research Ethics Committee (BM17/1/36) and the Western Cape Government Department of Health. In Sweden, this was the Regional Ethics Review Board in Stockholm (Ref. 2015/712-31/1). --- Consent for publication Not applicable. --- Competing interests The authors declare that they have no competing interests. • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year --- • At BMC, research is always in progress. --- Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research? Choose BMC and benefit from:? Choose B

Background: Understanding motivational determinants of physical activity (PA) is essential to guide the implementation of PA at individual and population level. Knowledge about the cross-cultural generalizability of these determinants is lacking and they have mostly been studied as separate factors. This study compares a motivational process model across samples from diverse populations with, or at risk of diabetes. Methods: Measurement invariance of barrier identified regulation, barrier self-efficacy and social support was assessed in a rural Ugandan sample (n = 712) and disadvantaged samples with high proportions of immigrants in urban South Africa (n = 566) and Sweden (n = 147). These motivational determinants were then compared through multigroup structural equation modeling.The studied motivational constructs showed scalar invariance. Latent mean levels of perceived social support and barrier self-efficacy were lower in South Africa and Sweden. Structural models (for different PA outcomes) were not consistent across settings except for the association between perceived social support and identified regulation. Identified regulation was only associated with vigorous PA in Uganda and with moderate PA in South Africa. The association between social support and PA outcomes ranged from weak to not significant and the association between self-efficacy and PA was not significant. Self-reported PA was highest in Uganda and lowest in Sweden. Selfreported vigorous PA was significantly related to lower hemoglobin A1c levels, while moderate PA was not. Conclusions: Findings suggest that: 1) it is feasible to compare a motivational process model across diverse settings; 2) there is lower perceived social support and self-efficacy in the urban, migrant samples; 3) identified regulation is a more promising determinant of PA than self-efficacy or social support in these populations; 4) associations between motivational determinants and PA depend on the perceived type and/or intensity of PA; 5) perceived relatedness functions as a basic psychological need across diverse settings; and 6) people's perception of the PA they perform depends on their perceived level of intensity of PA which would have major implications for health promotion.

study. The study was approved by the ethics committees in each of the respective countries. In Uganda, this was the Higher Degrees, Research and Ethics Committee (HDREC) of Makerere University School of Public Health and the Uganda National Council for Science and Technology (Ref. HDREC-331 and HS 1917 respectively). In South Africa this was the University of the Western Cape Biomedical Science Research Ethics Committee (BM17/1/36) and the Western Cape Government Department of Health. In Sweden, this was the Regional Ethics Review Board in Stockholm (Ref. 2015/712-31/1). --- Consent for publication Not applicable. --- Competing interests The authors declare that they have no competing interests. • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year --- • At BMC, research is always in progress. --- Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research? Choose BMC and benefit from:? Choose BMC and benefit from: --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: Understanding motivational determinants of physical activity (PA) is essential to guide the implementation of PA at individual and population level. Knowledge about the cross-cultural generalizability of these determinants is lacking and they have mostly been studied as separate factors. This study compares a motivational process model across samples from diverse populations with, or at risk of diabetes. Methods: Measurement invariance of barrier identified regulation, barrier self-efficacy and social support was assessed in a rural Ugandan sample (n = 712) and disadvantaged samples with high proportions of immigrants in urban South Africa (n = 566) and Sweden (n = 147). These motivational determinants were then compared through multigroup structural equation modeling.The studied motivational constructs showed scalar invariance. Latent mean levels of perceived social support and barrier self-efficacy were lower in South Africa and Sweden. Structural models (for different PA outcomes) were not consistent across settings except for the association between perceived social support and identified regulation. Identified regulation was only associated with vigorous PA in Uganda and with moderate PA in South Africa. The association between social support and PA outcomes ranged from weak to not significant and the association between self-efficacy and PA was not significant. Self-reported PA was highest in Uganda and lowest in Sweden. Selfreported vigorous PA was significantly related to lower hemoglobin A1c levels, while moderate PA was not. Conclusions: Findings suggest that: 1) it is feasible to compare a motivational process model across diverse settings; 2) there is lower perceived social support and self-efficacy in the urban, migrant samples; 3) identified regulation is a more promising determinant of PA than self-efficacy or social support in these populations; 4) associations between motivational determinants and PA depend on the perceived type and/or intensity of PA; 5) perceived relatedness functions as a basic psychological need across diverse settings; and 6) people's perception of the PA they perform depends on their perceived level of intensity of PA which would have major implications for health promotion.

Introduction --- A recent meta-analysis of mainly cohort studies showed lower social-economic position as a consistent predictor of type 2 diabetes mellitus (T2D) incidence in high-and middle-income countries, using well-known individual socioeconomic indicators. 1 This and other major articles emphasised the lowest vs. highest social grouping; rather than summarising social gradient throughout all society stratification. 2,3 Meanwhile, several sex-specific inconsistencies have arisen from well-performed prospective studies on common social indicators and T2D risk, with weak or non-existent associations among men, 2,[4][5][6][7] but mostly positive associations among women, 4,8 although not in all. 2 The mentioned articles as well as others have focussed on different life-span stages, mainly in childhood and adulthood, but less information is available on a life-course approach to socioeconomic position models 9 with T2D incidence. 2,4,8,10 Social position does not have a direct biological effect on T2D. 11 Instead its effects are mediated by other risk factors that can be biologically related to the social patterning of the disease. Physical inactivity, obesity, cigarette smoking and unhealthy diet are considered potentially important mediators of the association between social position and T2D risk. 12,13 Recent evidence on life-long social position has established the role of obesity as the principal mediator on social position disparities, and a partial contribution of smoking. [13][14][15] Our objective was to assess the association of individual lifecourse social position (LiSoP) with T2D risk, and its mediators influence, among adult men and women from the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort in Spain. --- Methods --- Study sample The EPIC research project is a large prospective and ongoing study involving a population from 10 European countries. Details of the methodology employed in the EPIC study have been published previously. 16,17 The EPIC-Spain cohort comprised 41 438 participants, mostly aged 30-65 at the time of enrolment (1992-1996), and recruited among healthy volunteers, including blood donors, civil servants and the general population. Extensive self-reported questionnaire information on diet, lifestyles and medical history was collected at baseline, and anthropometric variables were measured according to standard procedures. The cohort covers a diverse range of sociooccupational levels and territorial idiosyncrasies of five Spanish provinces from the North/Atlantic Ocean (Asturias, Guipu <unk>zcoa and Navarra) and the South/Mediterranean Sea (Murcia and Granada) environments. All participants gave their informed consent, and a Medical Ethical Review Board (Bellvitge Hospital, Barcelona) granted approval to the project. --- Identification of incident T2D cases T2D incident cases were ascertained and verified between recruitment and 31 December 2006 with a mean follow-up time of 12.1 years (AE1.8 SD). Ascertaining T2D cases was based on different sources of information, including self-reported diabetes or consumption of diabetes medication during a follow-up interview 3 years after recruitment; hospital discharge databases; drug prescription records; regional mortality registers and the National Death Index; and record linkage with primary care registers. Access to laboratory data on glycaemia and glycosilated haemoglobin (HbA1c) tests were available in Guipu <unk>zcoa. Possible T2D cases were ascertained by trained health professionals through careful review of clinical data and health information available. Definite cases were defined by physician's diagnosis of T2D present in the medical history, or otherwise evidence of diabetes from two independent sources: 2-h post-load glycaemia value!200 mg dl <unk>1 after a 75-g oral glucose tolerance test, HbA1c >7%, fasting plasma glucose!126 mg dl <unk>1, non-fasting glycaemia!200 mg dl <unk>1, diabetes-related medical visit or medical death certificate (code E11 of The Tenth International Statistical Classification of Diseases and Related Health Problems, http://apps. who.int/classifications/icd10), self-report of diabetes, use of antidiabetic medication (A10 code of the World Health Organization Anatomical Therapeutic Chemical classification system (www.whocc.no). The ascertainment and verification process of T2D cases followed the criteria and procedures defined in the EPIC-InterAct study. 18 The incidence date was defined as either the earliest date of diagnosis or first antidiabetic drug use registered in medical records, or the date of self-report. --- Assessment of LiSoP Data on education and paternal occupation were gathered by trained personnel during a face-to-face interview at enrolment. Participants were asked about their highest educational qualification attained in adulthood, and grouped according to the International Standard Classification of Education (ISCED-11) (www.uis.unesco.org) into: non-formal education; primary school; technical training; secondary school; and university degree or higher. The childhood social position based on father's occupation when the participant was 10 years old was registered and coded into the national version (CNO-94) of the International Standard Classification of Occupations (ISCO88), and further grouped for standard epidemiological analysis and presentation 19 as: I (managers, administrators and professionals, all higher-grade); II (other managers or administrators and professionals, and technicians); III (self-employed and staff services sector and supervisor workers); IV (semiskilled/skilled manual workers); and V (unskilled manual workers). Fewer than 6% of the cohort had missing data on exposure variables. A hierarchical combination of childhood and adulthood social position was then computed in order to assess the individual life-course social risk accumulation under a Weberian approach, 9,20,21 as the points for own adult education (university = 0, secondary = 1, technical = 2, primary = 3, non-primary = 4 points) plus childhood paternal occupational category (I = 0, II = 1, III = 2, IV = 3, V = 4 points). The score was categorised into: upper (0-2 points), middle (3-5 points) and lower (6-8 points) classes. 22 Anthropometric measurements, dietary history, lifestyleand clinical information Height (cm), weight (kg) and waist and hip circumference (in cm) at baseline were measured during a physical examination according to standard methodology. Waist circumference (WC) was categorised into 'normal' (<unk>102 cm in men, or <unk>88 cm in women) and 'high' groups (!102 cm in men, or!88 cm in women) according to NCEP/ATPIII criteria. 23 Body mass index (BMI) was obtained as weight divided by square height (in kg m <unk>2 ) and grouped into WHO standard categories of normal weight (<unk>25 kg m <unk>2 ), overweight (25 to <unk>30 kg m <unk>2 ) and obesity (!30 kg m <unk>2 ). Missing anthropometric values (1%) were replaced by single imputation including sex, age and education as predictors in multiple regression models. Information on the previous year's habitual diet was gathered by means of a validated dietary history method during a personal interview. 24,25 Energy and nutrient intakes were calculated using a specific food composition table. 26 Correction for misreporting energy intake was considered, classifying participants as underreporters, plausible reporters and over-reporters according to the predicted total energy expenditure method. 27 Time spent in domain-specific physical activity (PA) was assessed at baseline using a brief validated questionnaire. 28 A continuous variable of non-occupational PA was obtained as the sum of recreational and household activities, expressed in MET-h week <unk>1. PA at work was registered by asking participants to classify their job as sedentary, standing occupation, manual work or heavy manual work, among workers. Cigarette smoking status was defined as: never smoker; former smoker for 10 years or more; former smoker for fewer than 10 years; current smoker of up to 10 cigarettes day <unk>1 ; current smoker of 11-20 cigarettes day <unk>1 ; current smoker of more than 20 cigarettes day <unk>1 ; or unknown. Age when they starting smoking was also recorded and accounted for in the analyses. Finally, the questionnaire gathered self-reported clinical information on the presence of hypertension, hyperlipidaemia, cancer or cardiovascular disease (yes/no/unknown). Women were further asked about ever use of oral contraceptives or hormonal replacement therapy, and reproductive history. --- Statistical analyses For the present analysis, exclusions were applied for prevalent T2D cases (n = 2383), non-type 2 diabetics (n = 4) and participants with unknown diabetes status or diagnosis date (n = 713). Among the remaining participants, those with missing data on childhood paternal occupation (n = 359) or own adult education (n = 1622) were further excluded, leaving 2357 incident cases of T2D, 36 296 participants and over 410 000 person-years available for analysis. Descriptive statistics of participants' baseline characteristics based on means and standard deviations or frequencies and percentages (as appropriate) were presented for men and women separately, and stratified by LiSoP. The Kruskall-Wallis and 2 tests analyses were used to compare continuous and categorised baseline variables across LiSoP groupings, respectively. Hazard ratios (HR) and 95% confidence intervals (95% CI) of incident T2D risk by levels of social position were calculated using Cox proportional hazards regression, fitted separately by sex. All models were stratified by centre (to account for differences in recruitment and follow_up procedures) and age at recruitment (in 1-year categories). Age was the underlying time variable, with entry time defined as age at recruitment, and exit time as age at the date of the T2D diagnosis, emigration, death or end of follow-up, whichever occurred first. The proportionality assumptions were tested on the basis of Schoenfeld residuals. A basic stratified model was defined for LiSoP as predictor. Subsequently, several block models were added individually: Anthropometry (BMI, WC and hip circumference); Smoking (current cigarette use, and age at smoking initiation); Diet and PA (total energy intake [kcal day <unk>1 ], plausibility of energy intake reporting, macronutrient intake-proportion of energy from protein, lipids and carbohydrates-, alcohol consumption [in g day <unk>1 ], daily intake [g day <unk>1 ] of vegetables, fruit, red meat, processed meat and coffee, plus PA at work and during leisure time in [MET-h week <unk>1 ]); Chronic Disease (prevalent hypertension or hyperlipidaemia at recruitment, or history of cancer or cardiovascular disease); Reproductive Factors (postmenopausal status, ever use of oral contraceptives or hormonal replacement therapy). Lastly, all variables previously considered were mutually adjusted in final models. The upper class was always defined as the reference category. Dietary, anthropometric and leisure-time PA variables were modelled as continuous. Restricted cubic spline transformations were applied to anthropometric variables (except height) and alcohol to account for their non-linear relationships with T2D risk, in order to minimise residual confounding in LiSoP associations with T2D. Only total energy intake and PA variables were kept in the models, irrespective of their statistical significance. Potential effect modification of the association between LiSoP and T2D risk was tested for a set of variables including BMI, WC, smoking status (adjusted by age of initiation), age (<unk>50 and!50 years) and baseline chronic disease. Effect modification was tested by comparing models evaluating the interaction of LiSoP (three categories) with models without the interaction terms, using likelihood ratio tests. For stratified analyses based on weight categories defined by either BMI or WC, a parsimonious set of adjustment variables was selected based on their statistical significance. Sensitivity analyses performed by excluding participants with fewer than 2 years of follow-up or those with missing anthropometric showed no relevant differences with main models. Analyses were performed with STATA/SE version 10.1 (STATA Corp., College Station, TX). All P-values were two-sided and evaluated at the 5% level of statistical significance. --- Results Overall baseline participant characteristics showed significant differences across LiSoP categories, except for hip circumference, daily energy from proteins and carbohydrates, and fruit intake, in men; and for energy intake from protein, in women (table 1). After the follow-up period, age-adjusted incidence rates of T2D increased as LiSoP groups lowered among men and women (Supplementary table S1). Table 2 shows HR for T2D of LiSoP classes and social gradient from different Cox models stratified by centre and age, and adjusting for specific grouping of mediators (see models in 'Methods' section). Adjusted HRs from final models evidenced significant risks excess of 64% for middle and 90% for lower classes, in men; and of 104 and 126%, respectively, for these two classes, in women. Concurrently, LiSoP presented significant social gradients into T2D risk (P <unk> 0.01) for T2D in both sexes. Taking the different LiSoP and T2D incidence models into consideration anthropometry was shown to be the main input, performing almost as closely to the mutually adjusted models for both men and women. The other main explanatory models regarding smoking, diet and PA, chronic disease and reproductive factors obtained similar HRs in the respective class and sex to the basic model of LiSoP (table 2). 1). The evaluation of effect modification (table 3) showed a significant interaction in women for WC (normal vs. excess) and smoking status (never/former/smokers), and social gradients for the normal WC and current smokers. Men presented higher HRs for T2D in excess weight, never smokers and <unk>50 years old, as compared with their counterpart risk groupings, whereas HRs were higher among women without previous chronic disease. Furthermore, stratification of standard BMI categories by WC groups consistently revealed higher T2D risks at lower LiSoP categories among men and women, yet the limited number of cases in some subgroups may have restricted the power to attain the level of statistical significance (Supplementary table S2). --- Discussion The results strongly suggested that LiSoP was significantly related to T2D risk in Spanish adult men and women. Obesity would be the major biological mediator driving this relationship. WC in men and women, and BMI in men, have specifically been shown as the main explanatory anthropometric mediators, in agreement with previous results on the anthropometric indices of T2D in the EPIC-Spain cohort. 29 The strengths of this study include the large sample size, the prospective design and the validation of T2D cases, thus reducing the potential for misclassification bias. Moreover, a wide set of variables of the behavioural pathway for diabetes outcomes were available for analysis, such as measured anthropometric data; PA; detailed dietary information; alcohol intake; and smoking status and intensity; as well as reproductive status and hormonal drug therapy in women. Therefore, relevant models of the relationship between social indicators and T2D could be evaluated; these accounted for different types of mediators; thus, providing insight into the underlying social pathways of T2D. Nevertheless, this study also has limitations. Those arising from the definition of the LiSoP score were previously explained in parallel to most Western countries official socioeconomic classifications, 22 thus supporting the external validity of the reported associations. Further, education has been related to early adulthood and a suitable predictor for occupation. 20,30 In addition, childhood anthropometrics were not measured. Increased body size starting from childhood was associated with a greater risk of diabetes in adult women, unless they become lean in adulthood. 31 Adult social position and other obesity determinants may be the mechanisms responsible for the observed associations between childhood and adult obesity. In that sense, a systematic literature review (years 1998-2008) suggested that childhood social position was inversely related to adulthood obesity in women and not associated in men. 32 Moreover, a recent systematic review (2008-2010) showed childhood socioeconomic position to be associated with T2D and obesity in adulthood. 14 In that way, BMI measurement at baseline could be a mediator of childhood social position, which attenuated the T2D and LiSoP relationship towards non-statistical significance. Data on family history of diabetes were not available. This could mean that more cases with low social position were diagnosed during follow-up and hence, the results may be overestimated. The lack of data pertaining to mental diseases might be regarded as another limitation. In that sense, it has been argued that the relationship between mental depression and T2D is bi-directional, with a relevant risk of T2D, and a modest increase on the contrary. 33 We acknowledge criticism of the standard methodology applied to evaluate the intermediary effect of potential mediators, as it is susceptible to possible bias. 34,35 However, alternative methods for identifying biological mediation are not widespread and may need to be further developed to overcome their own limitations. Finally, as in most longitudinal studies, data on potential confounders were only available at baseline, so it was therefore not possible to control for differential exposures to T2D-related factors at different follow-up times. Our results on LiSoP and T2D concur with prospective evidence from other cohort studies in diverse international settings. 2,8,36 All studies have observed HRs lower than in our Spanish cohort, although they are barely comparable to US racial and access constraints to the health care system. Regarding gender-specific differences, we have found social inconsistency effects among women, in accordance with other prospective studies, 37 and at the same time, we confirmed an earlier reported social gradient effect on T2D risk among men. 6 Most of the articles published have shown BMI as a major single mediator, in concordance with our results. 13,37,38 Others have described the importance of models including WC, low social position in childhood and adulthood weight excess, especially in women. 36 Our differences on smoking relevancy in contrast to other studies could be related to the measurement types used, such as for smoking status or smoking history categories, are probably valid. 13,15 Anyway, due to effect modification, current women cigarette smoking should be taken into account to close the social-and gender-gap in governmental health policies on weight excess reduction. 38 In the European context, self-reported T2D was significantly related to education among women, after adjusting for age, BMI, alcohol consumption, smoking status and physical inactivity, in 11 countries. That relationship was principally mediated by BMI. 39 Table 3 HR a and 95% confidence intervals (CI) of T2D in men and women, stratified by potential effect modifiers according to life-course social position: The EPIC-Spain Cohort Repeated measurements after recruitment in cohort participants' would document the changing pattern of diet, PA, anthropometry and other related health behaviours. Meanwhile, social position after retirement would allow a more accurate estimation of T2D risk, as cohort participants were mostly elderly. It would be desirable if future analyses of this EPIC cohort included the aforementioned issues in relation to T2D risk. LiSoP was related to T2D risk in adult Spanish men and women. Anthropometry mostly explained the relationship between LiSoP and T2D. WC and body mass indices were the major mediators from LiSoP to T2D risk. As WC, BMI and even LiSoP were partly avoidable and modifiable, this offers a great opportunity for community health action to target and counterbalance adverse social strata and health disparities. --- Supplementary data Supplementary data are available at EURPUB online. --- Conflicts of interest None declared. --- Key points Life-course social (LiSoP) gradient was established through major social strata for diabetes mellitus type 2 (T2D) risk in adult men and women in Spain. Obesity was confirmed as the major biological mediator driving the relationship between individual LiSoP and T2D in a Southern European context. Waist circumference in both sexes and body mass index in men have been revealed as the main explanatory anthropometric indices for obesity in the relationship between T2D risk and life-long social position exposure. Background: Health inequalities have increased over the last 30 years. Our goal was to investigate the relationship between low individual socioeconomic status and poor breast cancer prognosis. Our hypothesis was: low socioeconomic status patients have a higher risk of being diagnosed with late stage breast cancer than high socioeconomic status ones due to delayed diagnosis. Methods: We conducted a matched case-control study on 619 women with breast cancer, living in the He <unk>rault, a French administrative area. Both Cases and Controls were recruited among invasive cases diagnosed in 2011 and 2012 and treated in He <unk>rault care centers. Cases were defined as patients with advanced stages. Controls were composed of early stage patients. Individual socioeconomic status was assessed using a validated individual score adapted to the French population and health care system. Results: We observed that low socioeconomic status patients have a 2-fold risk of having late stage breast cancer regardless of cancer characteristics and detection mode (screening vs. clinical signs). Conclusion: One reason explaining those results could be that low socioeconomic status patients have less regular follow-up which can lead to later and poorer diagnosis. Follow-up is improved for women with a better awareness of breast cancer. Health policy makers could reduce health inequalities by reducing the delay in breast cancer diagnosis for low socioeconomic status women. --- Introduction --- B reast cancer (BC) is the most prevalent cancer among women. In France, almost 50 000 new cases were diagnosed in 2012. 1 With a 10-year survival rate of 76% 2 BC is increasingly associated with positive prognosis. In France, over the past 30 years, the mortality rate of BC patients has slightly decreased 1 due to an improvement in treatment and medical care. At the same time, the proportion of early stage BC often detected by mammography screening (MS) has increased. Since the end of the 1980s, socioeconomic factors such as socioeconomic status (SES) have been investigated as prognostic factors of BC. [3][4][5] An association between low SES and BC patient survival has been observed in international studies. Low SES women with BC had a lower survival rate than high SES women. [6][7][8][9] Differences in survival rates for patients with BC have been observed as being dependent on SES. The survival rate of high SES populations was found 10% higher than among low SES populations. 7 However, the mechanisms of the association between BC survival and SES remain unconfirmed and the process itself is not yet clear. Such survival rate differences could be partly explained by'stage at diagnosis'. 10 Indeed, studies have found that low SES populations had a higher risk of advanced diagnostic stages at discovery, [11][12][13][14][15][16][17][18][19] probably due to delayed diagnosis. In France, a positive association between low SES and late diagnostic stages has been described. 18,19 Thus, our research aimed to strengthen this

Background: The literature has consistently shown that extreme social-economic groups predicted type 2 diabetes mellitus (T2D), rather than summarising the social gradient throughout all society stratification. Body mass index (BMI) was established as the principal mediator, with little support for other anthropometries. Our aim was to investigate an individual life-course social position (LiSoP) gradient and its mediators with T2D risk in the EPIC-Spain cohort. Methods: 36 296 participants (62% women), mostly aged 30-65 years, and free of T2D at baseline (1992-1996) were followed up for a mean of 12.1 years. A combined score of paternal occupation in childhood and own adult education assessed individual life-course social risk accumulation. Hazard ratios of T2D were estimated using Cox regression, stratifying by centre and age, and adjusting for different explanatory models, including anthropometric indices; dietary history; smoking and physical activity lifestyles; and clinical information. Results: Final models evidenced significant risks in excess of 63% for middle and 90% for lower classes of LiSoP in men; and of 104 and 126%, respectively, in women. Concurrently, LiSoP presented significant social gradients for T2D risk (P < 0.01) in both sexes. Waist circumference (WC) accounted for most of the risk excess in women, and BMI and WC in men. Conclusions: LiSoP gradient was related to T2D risk in Spanish men and women. WC mostly explained the relationship in both genders, together with BMI in men, yet LiSoP retained an independent effect in final models.

Introduction Health is important for a variety reasons, in particular for individuals' well-being and ability to pursue different life plans [1]. However, health is quite complex as, according to the World Health Organization, it is "a state of complete physical, mental and social well-being" [2]. The complexity of health is reflected by the production function of health, which was first described by Auster, Leveson and Sarachek (1969) [3]. They examined health (measured by mortality rate) as a function of both medical care and environmental variables [3]. Thus, the health production function describes "the relationship between combinations of medical and non-medical inputs and the resulting output" [4]. Many researchers have used this concept in their studies, but they have also started to employ different variables to explain health status [5]. Thus, many empirical analyses have considered income as one of the main determinants of health, followed by education, food quality, health expenditure, social protection, early childhood development, food insecurity, unemployment and job insecurity, working life conditions, housing, basic amenities and environment, social inclusion and non-discrimination, structural conflict and access to affordable health services of a decent quality [6]. Table 1-based on selected existing studies-shows the diversity of the determinants of health and health outcome measures in use. These analysed variables are actually the conditions in which people are born, grow up, live, work and age; in the literature they are described as the socioeconomic determinants of health [7,8]. --- of 20 The socioeconomic determinants of health can be grouped into five key areas or domains [9]: (1) economic stability; (2) education access and quality; (3) neighbourhood and built environment; (4) social and community context; and (5) health care access and quality. The first group of socioeconomic determinants applies to economic stability and social status. It includes working conditions, which can provide financial security, social or employment status, social relations, personal development, self-esteem and protection from physical and psychosocial hazards [9]. Unemployment can cause psychosocial stress and can thus have a strong impact on physical and mental health and well-being [10]. In addition, as children age, the relationship between household income and their health becomes more pronounced [11]. Education (access and quality) is the second group of socioeconomic determinants. It includes such factors as education attainment in general, early childhood education and development, high school graduation, enrolment in higher education, language and literacy [7,12]. The positive relationship between more years of education and better health is one of the most fundamental connections in health economics, which originates from Grossman's health demand model [13]. In addition, it is an empirically well-documented pattern in the literature [14][15][16]. Education improves health, as it increases knowledge, skills, reasoning, effectiveness and other abilities as well as enhancing a sense of personal control that can enable a healthy lifestyle [17][18][19]. The third group of socioeconomic determinants is the quality of one's neighbourhood and built environment (physical environment). This group includes housing and shelter, transportation and roads, food, air and water quality, sanitation, neighbourhood crime and violence, safety, parks, walkability, and local geography and communities [10]. Homelessness and overcrowding are risk factors of physical and mental health [8,10]. The current model of urbanisation and the lack of balance between rural and urban areas pose many environmental challenges, especially those connected with climate change [12]. Good quality housing requires appropriate heating. Housing with insufficient heat is linked with a number of cardiovascular and respiratory morbidity outcomes, as well as increased incidence of psychological conditions such as depression. Heating also determines the hygrothermal conditions in the house. Inappropriate hygrothermal conditions may favour the proliferation of mites and asthma [20][21][22][23]. By having access to energy, people can have vital health determinants in the house, such as heating and the means to prepare nutritious food [24]. In addition, many empirical studies have shown that people's perception of the built environment may directly influence mental stress, which has an impact on their well-being and overall health. It can also indirectly influence health through health-related behaviours such as social contact or physical activity [25]. Additionally, active transportation can increase physical activity and thus health [26,27]. The next group of socioeconomic determinants-social and community context (social support networks)-includes community engagement, social integration, support systems, the beliefs of the family and community, customs and traditions, civic participation and discrimination. Having greater support from families, friends and communities is linked to better health [9,28]. The final group of determinants applies to access to and quality of health care, which means access to primary health care, health insurance coverage, health literacy, quality of care and utilisation of health care; these elements are integral in the prevention and treatment of disease and generally influence health [10, 29,30]. By forming the conditions of daily life, these socioeconomic determinants influence people's opportunities to be and stay healthy [31]. Moreover, these socioeconomic determinants of health have varied across space and time over the last few decades. Since they are also under the influence of different socioeconomic processes, some of them are more significant or fundamental than others. For example, the process of urbanisation has improved standards of living, positively influencing health, but negatively impacting lifestyles and lifestyle-related factors, including physical inactivity and unhealthy diets [32]. In addition, deindustrialisation has led to greater levels of socioeconomic deprivation (and associated factors) and has resulted in relatively poor health status among people in deindustrialised areas [33]. The global financial crisis has caused a dramatic transformation of employment in the weakest economies of the Eurozone. The deterioration of working conditions, low pay and periods of prolonged unemployment for most of the working population-especially women-have been observed [34]. Financialisation has exerted significant effects on many aspects of our daily life, such as consumption patterns, housing affordability, employment structure and social conditions, which are relevant to health. Generally, it has contributed to increased income inequalities through different channels [35]. The uneven distribution of socioeconomic determinants contributes to intergroup differences in health outcomes, both within and between societies [31], which is a major obstacle in achieving health justice [36]. Thus, ensuring health equity requires the elimination of unfair and avoidable differences in health among population groups, which are defined economically, socially, geographically or demographically [37,38]. Thus, the unequal allocation of power as well as resources, which appears in unequal social, economic and physical conditions, is recognised as one of the main root causes of health inequity [7,39]. It is mainly derived from the existence of inequalities in other areas of life, such as economic, political or social spheres [40]. This is the result of decisionmaking processes, policies, social norms and structures, which exist at all levels in society; therefore, effective interventions are required in all sectors [31,41]. Thus, the socioeconomic determinants are modifiable and can be influenced by social, economic and political processes, historical and contemporary policies, law, investment, culture and norms [7]. As socioeconomic determinants affect how people experience the world and the choices they make, bringing about a reduction in their distribution inequities is an important challenge for health policies [42,43]. Therefore, a burgeoning volume of research is focused on the social, economic and environmental determinants of health and their impact on health outcomes, as well as identifying these determinants as the main root cause of many health inequities. Many studies-which have covered different world populations and various ranges of socioeconomic determinants (Table 1)-have shown that people from lower socioeconomic groups have shorter lives and more often suffer from health problems, while people with a quality education, stable employment, safe homes and neighbourhoods and access to preventive services tend to be healthier throughout their lives. Previous research in the area of socioeconomic inequities in the health of Poles has primarily compared Poland with other countries [44,45]. There are other studies having a limited context, such as those focusing on economic status and gender [46]; education, marital status, employment status or place of residence and their impact on mortality among working-age people [47]; the social determinants of the self-rated health of Polish women and men [48]; and the relation between expenses for health and healthy life expectancy [49]. There have also been a few studies on the socioeconomic determinants of the health of rural inhabitants [50,51]. As reducing health inequities is treated as a matter of social justice and is thus a kind of ethical imperative, the Commission on Social Determinants of Health called on the WHO and all governments "to lead global action on the social determinants of health with the aim of achieving health equity" [12]. In the Polish health care system, both health and equity are important values [52,53], as determined from the WHO constitution and strategy [2] as well as from Article 68 of the Polish Constitution [54]. In addition, one of the strategic objectives of the Polish national health policy, as formulated in the National Health Programme [55], is the elimination of geographical and social inequalities in health. Therefore, in this study we focus on identifying inequities in the distribution of socioeconomic determinants of health within Poland and the impact of socioeconomic determinants on the health status of Poles, as measured by mortality rate. The aim of this study is to measure the level of inequities in the distribution of socioeconomic determinants of health between geographically defined groups of people in Poland. The hypotheses are as follows: (1) There are inequities in the socioeconomic characteristics within geographically defined population groups. (2) Some socioeconomic determinants of health have a particularly strong impact on the health status of Poles. In order to verify these hypotheses, we used the database of Statistics Poland [56], which determined the final range of socioeconomic variables and the year of research (2018) adopted for the study. Therefore, it was possible to derive the data at the powiat level-the second (out of three) administrative level in Poland-which made it possible to identify the sources of geographical inequities. Self-rated health. Education, household income levels. Dunn, J.R., Walker, J.D., Graham, J., Weiss, C.B. [85]. Self-reported health. Social support, type of housing, gender. Miura, K., Takamori, A., Hamazaki, K., Tsuchida, A., Tanaka, T., Origasa, H., Inadera, H. [86]. Health-related quality of life. Dietary pattern. Kim, J.-H., Park, E.-Ch. [87]. Health-related quality of life, overall quality of life. Household income, education levels. Source: Authors' study. As it was conducted on the powiat level, our research fills an existing gap by providing more specific information on the spatial diversity of the Polish population in terms of the socioeconomic determinants of health. The novelty of this research also arises from it being the first time the two-stage nested Theil decomposition method is utilised in the context of the Polish population, allowing national inequity to be decomposed to macro-regions by comparing inequities between voivodeships and within voivodeships. --- Materials and Methods Poland is administratively divided into three levels, including 16 voivodeships (provinces), 380 powiats (including 66 cities with powiat status) and 2478 gminas (2018). Each voivodeship consists of powiats, and each powiat consists of gminas. According to the Nomenclature of Territorial Units for Statistics (NUTS), Poland is divided into 7 macro-regions (NUTS 1), then 17 regions (NUTS 2) and 73 subregions (NUTS 3). Each macro-region consists of regions, and each region consists of subregions. Voivodeships are conterminous with regions (NUTS 2), with one exception-Mazovian voivodeship-which is split into two NUTS 2 units (regions): Warsaw-capital and Masovian-regional. Thus, all (16) voivodeships can be classified into 7 macro-regions (NUTS 1). The above relations between macro-regions, regions and voivodeships in Poland are presented in the Table 2. Individual macro-regions reflect the economic and social development of various regions in Poland (Supplementary Material). --- Table 2. Relations between macro-regions, regions and voivodeships in Poland. --- Macro-Region (NUTS 1) Region (NUTS 2) Voivodeship South-West Lower-Silesian (DL) Opole (O) Lower-Silesian (DL) Opole (O) South Lesser Poland (M<unk>) Silesian ( <unk>L) Lesser Poland (M<unk>) Silesian ( <unk>L) North-West Lubuskie (LU) Wielkopolska (WL) Zachodniopomorskie (ZP) Lubuskie (LU) Wielkopolska (WL) Zachodniopomorskie (ZP) North Kuyavian-Pomeranianranian (K-P) Warmian-Masurian (W-M) Pomeranian (PO) Kuyavian-Pomeranianranian(K-P) Warmian-Masurian (W-M) Pomeranian (PO) Central <unk>ód <unk> (<unk>) <unk>wi<unk>tokrzyskie ( <unk>W) <unk>ód <unk> (<unk>) <unk>wi<unk>tokrzyskie ( <unk>W) East Lubelskie (LE) Podkarpackie (PK) Podlaskie (PL) Lubelskie (LE) Podkarpackie (PK) Podlaskie (PL) Masovian Warsaw-capital (W-Sto<unk>) Masovian-regional (MAZ_R) Masovian (MAZ) Source: Statistics Poland. Thus, the level of unit data is the level of powiats for the purpose of analysing the inequity at the level of macro-regions and then the 16 voivodeships. Data were derived from the Statistics Poland database for 2018 [56]. The scope of the obtained data covered 380 powiats in Poland-i.e., all of them-for the year 2018. Including all the powiats in Poland in the study made it possible to obtain representativeness in the research and the results. The range of variables and their measures (Table 3), i.e., socioeconomic determinants, were selected based on the analysis of previous studies ([ [18][19][20][21][22][23][24][25][26][27][28][29] and Table 1) and then determined by the availability of data. They were grouped into economic, education, employment, demography and built environment categories. The number of pupils enrolled in elementary school to the number of pupils who qualify for elementary education. --- EDJH-Gross scholarization ratios for junior high level The number of pupils enrolled in junior high school to the number of pupils who qualify for junior high school education. --- Employment EA-Employment rate in agriculture The percentage of the population aged 15-64 working in agriculture, forestry, hunting and fishing. --- EI-Employment rate in industry The percentage of the population aged 15-64 working in industry and construction. --- ES-Employment rate in services The percentage of the population aged 15-64 working in the trades, repairing of vehicles, transport and the warehouse industry, accommodation and catering and information and communication. --- EF-Employment rate in financial sector The percentage of the population aged 15-64 working in the financial and insurance sector and real estate market. --- UR-Unemployment rate The number of unemployed people as a percentage of the labour force. --- WAP-Working-age population The percentage of the working-age population. --- Demography FR-Feminization ratio Females per 100 males. --- OR-Old-age dependency ratio Population in the post-production age to 100 people of working age. --- Built environment WS-Water supply The percentage of people using the water supply system. --- SS-Sewage system The percentage of people using the sewage system. --- GS-Gas supply The percentage of people using a gas supply system. F-Forest area Forest area in hectares per capita. GR-Green area other than forest Green area (parks, lawns, etc.) in hectares per capita. DIS-Cultural buildings adapted for the disabled Cultural buildings adapted for the disabled per 1 square km. Source: Authors' study. First, we determined the descriptive statistics. Analysis of the average and median of the analysed health determinants (Table 4) suggests that in case of most of them (14 out of 17-IN, EDE, EDJH, EA, EI, ES, EF, UR, FR, OR, GS, F, GR, DIS), more than 50% of powiat values had levels lower than the average. Based on the standard deviation and variation, it can be found that the IN variable is characterized by a high level of dispersion. For the purposes of examining the distribution of the socioeconomic variables in Poland and to determine the drivers of inequity, the Theil index was employed. It was developed by Theil in 1967 and is widely used to measure spatial inequality [88]. The Theil index ranges between 0 and <unk>, where zero represents an equal distribution and any higher value represents a higher level of disproportion. Other commonly used methods to measure the level of inequity in the context of health and health care are the Gini index [38,52,59] and concentration index [57,58,60]. Compared with the Gini coefficient, when estimating regional differences, the Theil index allows sub-groups to be broken down within the context of larger groups. Thus, it is possible to analyse their contribution to the total differences and to identify the main sources of the overall differences [89]. This is an important property of the Theil index measure, as this additive decomposability implies that the aggregate inequality measure can be broken down into inequality within and between any defined population subgroups [90]. The main pitfall of the Theil index is that its values are not always comparable across completely different units, as in case of different nations. If the number and size of groups differ, then limit of the index will differ [91]. Since this article considered the three division scales of macro-region, voivodeship and powiat in Poland, it is more suitable to apply the two-stage nested Theil decomposition method as proposed by Takahiro Akita in 2003 [92]. This two-stage nested Theil index allows us to decompose the national overall inequality into between-macro-regions, between-voivodeships and within-voivodeships. Through such decomposition, the Theil index can comprehensively reflect the multi-scale inequality in the distribution of socioeconomic determinants, with each component explaining a part of overall inequality that is due to differences within and between voivodeships and between macro-regions. The overall national inequality, T, of a particular socioeconomic variable distribution, based on the powiat level, can be measured using the following formula [92,93]: T = <unk> i <unk> j <unk> k L ijk L log L ijk /L P ijk /P(1) where L ijk -the particular socioeconomic determinant of health in powiat k in voivodeship j in macro-region i; P ijk -the total population (or subpopulation, in case of EDE, EDJH, the children in the appropriate range of age were used as the subpopulation; in case of EA, EI, ES, EF, UR, the working age population was used as the subpopulation; where appropriate) in powiat k in voivodeship j in macro-region i; L-the overall national socioeconomic determinant of health; P-the overall national population (or subpopulation 1, where appropriate). Then, T ij is defined as the inequity in voivodeship j in macro-region i. T ij = <unk> k L ijk L ij log L ijk /L ij P ijk /P ij(2) T i, as the inequality in macro-region i, can be decomposed using the following equation: T i = <unk> j <unk> k L ijk L i log L ijk /L i P ijk /P i = <unk> j L ij L i T ij + <unk> j L ij L i log L ij /L i P ij /P i = T wi + T pi(3) where L ij -the particular socioeconomic determinant of health in voivodeship j and in macroregion i; P ij -the total national population (or subpopulation 1, where appropriate) in voivodeship j and in macro-region i; L i -the socioeconomic determinant of health in macro-region i; P i -the total national population (or subpopulation 1, where appropriate) in macro-region i; T wi -measures within-voivodeship inequality; T pi -measures between-voivodeships inequality. By combining all of the above formulas, the overall national differences, T, can be expressed as follows, which is the final form of the two-stage nested Theil decomposition method equation: T = <unk> i L i L T i + <unk> i L i L log L i /L P i /P = <unk> i L i L T i + T BR = <unk> i L i L T wi + T pi + T BR = <unk> i L i L T wi + <unk> i L i L T pi + T BR = T WP + T BP + T BR(4) where T WP -within-voivodeship component; T BP -between-voivodeship component; T BR -between-macro-region component. For the purposes of assessing the level of inequity of the analysed variables, the Herfindahl-Hirschman Index (HHI) [94] was also employed, which allows us to identify the level of inequity in the distribution of the socioeconomic determinants of health. It is commonly used in economics, health services research and other disciplines [95]. The HHI can be defined as the sum of square of the shares of each variable in the overall sum of variables, and it is expressed by the following formula [94]: HHI = <unk> n i = 1 (MS i ) 2(5) where MS i -the proportion of a percentage of a variable for i-powiats to a percentage of a variable in all powiats; n-number of powiats in the macro-region. The result is often multiplied then by 10,000; the distribution of variable is considered highly concentrated if the value of HHI is greater than 2500, moderately concentrated the HHI value is between 1500 and 2500, and unconcentrated if the HHI is between 100 and 1500 [96]. Then, multiple regression analysis was employed in order to identify the most significant determinants of health at the level of Polish powiats. Mortality rate was adopted as a measure of the health status of the population [97] and incorporated into the regression model as a dependent variable. The mean value of the dependent variable was 10.86, median 10.75, maximum 17.53, minimum 6.45, variance 2.33 and standard deviation 1.53. Distribution of the dependent variable was tested using the chi square test and was found to be normal. The Independent variables initially considered were the above-mentioned 17 determinants of health (see Table 3). In the first step of the analysis, two-sided correlations between each of independent variables and a dependent variable (the mortality rate) were assessed using the Spearman's rank correlation coefficient. Calculated absolute values of the coefficient are presented in Table 5. As a cut-off point of a significant correlation, the coefficient absolute value of 0.1 was adopted. Six of the independent variables (EDE, EDJH, EF, WS, F, GR) reached the absolute value of the correlation coefficient of less than 0.1 and were excluded from further analysis. The remaining 11 independent variables were included in the preliminary regression model. The second step of the regression analysis was development of the preliminary multiple linear regression model, containing the independent variables (IN, EA, EI, ES, UR, WAP, FR, OR, SS, GS, DIS) significantly correlated with the mortality rate. The general formula of the regression model is given below: Y = a 1 X 1 + a 2 X 2 +... + a n X n + B (6 ) where: Y-the predicted value of the dependent variable; X 1, X 2,... X n -the independent variables; a 1, a 2,... a n -the regression coefficients (slopes) of the independent variables; B-the intercept. The parameters (slopes and an intercept) of the preliminary model were established using the least squares estimation. For each of independent variables, a p-value was calculated employing the t-statistic. The significance level <unk> = 0.05 was adopted. a p-value above 0.05 indicated statistically non-significant variables. In the next step of the analysis, the preliminary model was refined. Four of the non-significant independent variables (IN, EI, ES, DIS) were excluded. The final model consisted of seven independent variables: EA, UR, WAP, FR, OR, SS, GS. The parameters of the final model (slopes and an intercept) were recalculated using the least squares approach, and t-statistics were employed for calculation of the independent variables' p-values as well. Additionally, the final model was tested with regard to statistical independence of the random errors with the use of the Durbin-Watson statistic. According to the D-W distribution tables, a value of the D-W statistic between 1.84 and 2.16 was adopted as an indicator of the absence of residual auto-correlations at a significance level <unk> = 0.05, which means that there is no violation of independence of the random errors in the final regression model. Calculations of the Theil index and the HHI were done using a free software spreadsheet. Calculation of the regression model was done using STATISTICA software (TIBCO Software Inc., Statistica version 13. (Palo Alto, CA, USA). --- Results The Theil index was employed to measure the nationwide equity of the distribution of socioeconomic variables in Poland and the contribution rate of each Polish voivodeship. The Theil index values shown in Table 6 indicate the existence of inequity in the distribution of such variables as GR, EF, F, EA, DIS, GS and ES. The values for these variables range from 0.1230-0.4644, while any value higher than 0 indicates some level of disproportion. In the case of the remaining variables, slight inequity can be observed, but the values are generally below 0.0684. It can be concluded that at the national level, individual areas in Poland vary in importance in terms of the size of green areas and forests. Some variation in the area of employment structure can be observed, as there is a concentration of employment in finance and agriculture as well as a slight concentration in services. Poland is also characterised by inequity in adapting buildings for people with disabilities and supplying gas to homes. Table 6 also contains results that show the contribution of three components of overall national inequality (T): the between-macro-region component (TBR), the betweenvoivodeship component (TBP) and the within-voivodeship component (TWP). In the case of the above-mentioned determinants (GR, EF, F, EA, DIS, GS and ERS), within-voivodeship inequity is largely responsible for their total unequal distribution, since the values of this component are generally higher than the other components (between-voivodeship inequity and between-macro-region inequity). However, the within-voivodeship component constitutes the main component of overall national inequities for socioeconomic determinants other than the education variables (EDE and EDJH) (see Table 6). In the case of the education variables, the differentiation between voivodeships is mainly responsible for the slight inequities at the national level. Thus, the difference within voivodeships is the main factor leading to national differences in the socioeconomic determinants of health distribution, from a spatial perspective. The results confirm the hypothesis that there is an inequality of the distribution of the socioeconomic determinants of health and that it is caused by within-voivodeship differentiation. As can be seen in Figure 1, the inequalities within voivodeships-i.e., between powiats-show different degrees of expansion, which led to the polarization of some of the socioeconomic determinants of health in 2018, such as forestation, gas supply, and the level of building adaptation for the disabled. High levels of inequity in the distribution of forests (F) were noted in the case of almost all voivodeships. Podlaskie (PL), Warmian-Masurian (W-M), Wielkopolska (WL) and Zachodniopomorskie (ZP) voivodeships presented some level of inequity in access to the gas supply system. Zachodniopomorskie (ZP) and Pomeranian (PO) voivodeships showed unequal distribution of cultural buildings adapted for the disabled.. Environ. Res. Public Health 2021, 18, x 12 o level of building adaptation for the disabled. High levels of inequity in the distribution forests (F) were noted in the case of almost all voivodeships. Podlaskie (PL), Warmia Masurian (W-M), Wielkopolska (WL) and Zachodniopomorskie (ZP) voivodeships p sented some level of inequity in access to the gas supply system. Zachodniopomors (ZP) and Pomeranian (PO) voivodeships showed unequal distribution of cultural bui ings adapted for the disabled. In the area of education and the labour market (Figure 2), it is noted that Lower Sile (DL) presented the highest inequity in the distribution of employees in the finance a services areas between powiats. This could be caused by the high concentration of finan and services companies in Wroclaw-the capital of Lower Silesia. The results present hi differentiation in Masovian (MAZ) and Silesian (<unk>L) voivodeships, as they show som level of concentration of both agricultural and finance employees. In the area of education and the labour market (Figure 2), it is noted that Lower Silesia (DL) presented the highest inequity in the distribution of employees in the finance and services areas between powiats. This could be caused by the high concentration of finance and services companies in Wroclaw-the capital of Lower Silesia. The results present high differentiation in Masovian (MAZ) and Silesian ( <unk>L) voivodeships, as they show some level of concentration of both agricultural and finance employees. These results suggest that these identified differences may be a capital city effect and may represent an urban-rural divide, which has been observed in other areas researched in Poland [98,99]. Populations continue to expand in and around many capital cities and urban areas, as they are associated with (perceived) education and/or employment opportunities. The HHI values for the socioeconomic determinants of health are presented in Table 7. The results present the concentration level of the above determinants and thus their distribution inequities. (DL) presented the highest inequity in the distribution of employees in the finance a services areas between powiats. This could be caused by the high concentration of finan and services companies in Wroclaw-the capital of Lower Silesia. The results present h differentiation in Masovian (MAZ) and Silesian (<unk>L) voivodeships, as they show so level of concentration of both agricultural and finance employees. Generally, the HHI values indicate a low level of variable concentration, as they are below 1500, especially in the case of the four macro-regions: south, north-west, north and east. If the HHI values are between 100 and 1500, then the particular feature is unconcentrated and is considered equally distributed. There is one exception, as the distribution of employment in the financial sector (EF) demonstrated moderate concentration in the south and north macro-regions (the values were between 1500 and 2500). The south-west and central macro-regions were characterised by moderate concentrations of most variables, apart from the employment rate in finance, which showed a high level of concentration (the values were greater than 2500). This high level of EF concentration, and such inequities in its distribution between macro-regions, may be due to the existence of large, fast-growing economic and financial city centres, such as Wroc<unk>aw (south-west) and <unk>ód<unk> (central). The Masovian macro-region was characterised by moderate concentration in the case of old-age dependency ratio and employment rate in both agriculture and industry (the values ranged from 1500 to 2500) and a high level of inequity in terms of the remaining socioeconomic variables (the values were greater than 2500). This may be due to the fast-growing capital of Poland, Warsaw, which is surrounded by relatively few developed areas. Initially, the 17 socioeconomic determinants of health listed in Table 3 were considered potential independent variables in a multiple linear regression analysis. Eleven of the determinants had a Spearman's rank correlation coefficient of over 0.1 and had sufficient two-sided correlation with mortality rate and were thus used for the construction of the preliminary regression model. These were IN, EA, EI, ES, UR, WAP, FR, OR, SS, GS and DIS. The parameters (slopes and an intercept) of the preliminary model are presented in Table 8. An r 2 value of 0.7906 indicates that this model describes about 79% of the variability in the mortality rate. Based on the t statistics, p-values for each independent variable were calculated; these are presented in Table 8. Five of the independent variables (IN, EA, EI, ES and DIS) were statistically nonsignificant (a p-value above 0.05). Four of them (IN, EI, ES and DIS) were excluded from further analysis, so the final regression model was constructed with seven independent variables: EA, UR, WAP, FR, OR, SS and GS. The independent variable

The purpose of this study is to identify inequities in the distribution of socioeconomic determinants of health within Poland and their impact on the health status of Poles, as measured by mortality rate. We hypothesised that (1) there are inequities in the socioeconomic characteristics within geographically defined population groups and (2) some socioeconomic determinants of health have a particularly strong impact on the health status of Poles. Poland is administratively divided into three levels: voivodeships, powiats and gminas. We used a dataset covering all 380 powiats in Poland for the year 2018. We employed a two-stage nested Theil index and Herfindahl-Hirschman Index. In order to identify which of these determinants has the strongest impact on health, we conducted a regression analysis. The study revealed some inequities in the distribution of socioeconomic determinants of health. The mortality rate can be partly understood from variations within voivodeships in the distribution of health determinants. Important national inequalities were found in the case of two determinants, which simultaneously proved their significant impact on Poles' health in the regression analysis. Thus, type of employment and access to modern infrastructure should be of particular concern for public authorities.

health listed in Table 3 were considered potential independent variables in a multiple linear regression analysis. Eleven of the determinants had a Spearman's rank correlation coefficient of over 0.1 and had sufficient two-sided correlation with mortality rate and were thus used for the construction of the preliminary regression model. These were IN, EA, EI, ES, UR, WAP, FR, OR, SS, GS and DIS. The parameters (slopes and an intercept) of the preliminary model are presented in Table 8. An r 2 value of 0.7906 indicates that this model describes about 79% of the variability in the mortality rate. Based on the t statistics, p-values for each independent variable were calculated; these are presented in Table 8. Five of the independent variables (IN, EA, EI, ES and DIS) were statistically nonsignificant (a p-value above 0.05). Four of them (IN, EI, ES and DIS) were excluded from further analysis, so the final regression model was constructed with seven independent variables: EA, UR, WAP, FR, OR, SS and GS. The independent variable slopes and the intercept in the final model are presented in Table 8. The r 2 value for the final model was 0.7220, which means that this model describes about 72% of the variability of the dependent variable (the mortality rate). In the final model, the p-value for all independent variables was below 0.05, indicating their statistical significance. A test of statistical independence of the random errors in the final model was done with the use of the Durbin-Watson statistic. The obtained value of the D-W statistic was 1.86, which means that at a significance level <unk> = 0.05, no residual auto-correlations occurred, and thus the model has good diagnostic features. --- Discussion This study identified inequities in the distribution of socioeconomic determinants of health between geographically defined populations. It demonstrates that in Poland, as a result of their geographic status, people do not have equal opportunity to achieve their full health potential. The results confirmed that voivodeships are quite heterogeneous in terms of the distribution of the socioeconomic determinants of health. This implies the existence of inequities in the distribution of these determinants. The main risk factors of health inequity are observed in the conditions of the built environment and employment. Discrepancies in access to green areas, forests and the gas supply system, as well as the levels of employment in agriculture and finance, were found. The selection of the powiat-level unit and the two-stage Theil index method allowed the identification of the level of national inequality in the distribution of the socioeconomic determinants of health in Poland. Moreover, these findings showed that this inequality across the country and in all macro-regions was decomposable and that the inequalities within voivodeships also represent an important part of national inequalities. The research only partly confirmed the existence of a high disproportion between eastern Poland (colloquially called Poland B) and western Poland (Poland A), which was recognised in other studies [100]. The within-voivodeship component constitutes the main component of overall national inequities, while the between-voivodeship component is only responsible for some national inequity in the case of the education variables (EDE and EDJH). In addition, the most diversified voivodeships are located in both the east and west of Poland. In addition, the values of HHI revealed that one macro-region in Poland-Masoviawas characterised by a high concentration of most of the health determinants. The Masovian macro-region contains the capital city, and such variation in the distribution of socioeconomic variables could be caused by different rates of development, leading to the growth of large centres and to increasingly poor surrounding areas, where there is no rapid economic growth [101]. Likewise, the south-west and central macro-regions were characterised by moderate concentrations of most variables. When we compare this result with those of Ucieklak-Je <unk> and Bem [51], who found that rural areas were homogeneous in terms of the analysed sociodemographic determinants of health, we suspect that the concentration of particular health determinants in urban areas could also have been the main reason for the variability among voivodeships or macro-regions. However, further research is required, as Ucieklak-Je <unk> and Bem [51] employed slightly different ranges of health determinants. The low level of most socioeconomic variable concentrations, which was recognised in the case of the south, north-west and north macro-regions, can be explained by historical factors, which many publications have described as a mechanism that still maintains regional disparity in Poland [102,103]. The period of partitions, in particular, contributed to differences in socioeconomic development and social resources in individual regions in Poland that still exist today. This period contributed to the diversification of the behavioural characteristics of the population of the particular partitions. Historical factors, therefore, caused regional differentiation in the importance awarded to local ties and economic attitudes [104], which, today, could favour equality or eliminate inequalities. The populations of the north and north-west macro-regions are characterised by greater entrepreneurship and a rational approach as well as greater economic activity, while the south of Poland is characterised by a high level of localism. As the northern and southern parts of Poland demonstrate similar levels of concentration in most socioeconomic determinants-lower than those of the other macro-regions-these results cannot be explained by variability in epidemiology [105]. Further research is required. Based on correlation and multiple regression analysis, only some of the 17 socioeconomic determinants of health taken into consideration proved to have a significant impact on the mortality rate of the Polish population. Six of the independent variables (EDE, EDJH, EF, WS, F and GR) were weakly correlated with the mortality rate (the absolute values of the correlation coefficient were less than 0.1). The remaining 11 independent variables (IN, EA, EI, ES, UR, WAP, FR, OR, SS, GS and DIS) were had significant two-sided correlation with the mortality rate and were used in the preliminary regression model. This model showed good predictive value and explained about 79% of the variability in the mortality rate. Nevertheless, not all independent variables in the preliminary model were statistically significant. The p-values calculated for four variables (IN, EI, ES and DIS) were much higher than the adopted <unk> = 0.05 (0.482, 0.561, 0.685 and 0.553, respectively), which means that their potential ability to predict the mortality rate value is uncertain, despite being sufficiently correlated with the dependent variable. In order to improve the regression model, these four variables were excluded from the final model. The final regression model consisted of seven independent variables: EA, UR, WAP, FR, OR, SS and GS. This model explained about 72% of the variability in the mortality rate, which is slightly less than in the preliminary model, but still represents good predictive value. For all independent variables, the p-values were less than 0.05, and their impact on the mortality rate could be perceived as being statistically significant. Based on the final regression model, four of the socioeconomic health determinants that were used had a positive influence on health status (they had negative regression slopes) and reduced the mortality rate: EA, FR, SS and GS. Three of the independent variables in the final model (UR, WAP and OR) had positive slopes. They increased the mortality rate and could be treated as risk factors of a deterioration in health status. In particular, the positive correlation between WAP and an increased mortality rate in the regression model requires further, focussed research. The within-country inequalities among these seven significant socioeconomic determinants of health identified in the Polish population could be particularly important to explain potential differences in health status at the powiat level. In the case of two significant determinants (EA and GS), the Theil index analysis indicated important national inequalities. These two determinants should not be interpreted too literally. EA can be treated more as an indicator of employment type (such as work in a healthy environment near one's residence that lacks strong subordination in the chain of command), while GS can be seen as an estimator of infrastructure development (such as modern infrastructure with no significant negative impact on the human environment and health due to low dust emission). These results could mean that socioeconomic determinants related to employment type and infrastructure development should be of special concern in improving the health status equity of the Polish population, inducing actions to facilitate equal access to modern ecological infrastructure and to make an active workforce market policy that prioritizes equal access to jobs without consequences for workers' health. The study led to the identification of the voivodeships that suffer the most from internal differentiation in the distribution of the socioeconomic determinants of health. In the case of access to gas supply, the Podlaskie (PL), Warmian-Masurian (W-M), Wielkopolska (WL) and Zachodniopomorskie (ZP) voivodeships presented some level of inequity. In addition, high differentiation between the Masovian (MAZ) and Silesian ( <unk>L) voivodeships was observed, as concentrations of both agricultural and financial employees were found. Thus, studies similar to this one could be used to support policymakers and local governments as well as other stakeholders responsible for creating public regional policy. Because many of these health differences are caused by decision-making processes, policies, social norms and structures, which exist at all levels in society, these results show the direction of changes that should be undertaken, especially in the Masovian macroregion. This study reveals that analysing variations in inequalities in the distribution of socioeconomic determinants of health within a country can help to identify entry points for policy. In this study, we proposed the two-stage nested Theil index to measure inequities in the socioeconomic determinants in Poland. This allowed analysis to be made at different statistical and administrative levels. --- Conclusions By using a dataset that covers all macro-regions in Poland in 2018, using the two-stage nested Theil index and conducting regression analysis, our results suggest that mortality rate (as an estimator of a population's health status) can be understood, in part, as the product of within-country variations in the distribution of inequalities of socioeconomic variables. These findings provide new evidence in this area, which is a current and developing global topic, and can add supporting arguments in the discussion of the future shape of social and health policy. This study contributes to science in a few ways. We provide new evidence in the area of socioeconomic determinants of health, underlying the importance of the health inequities as a result of unequal distribution of the gas supply and employment in agriculture. We also propose the use of the two-stage nested Theil index for inequity measures of the socioeconomic determinants of health in Poland. The limitations of the research arise from the range of available data. It would be valuable for Statistics Poland to collect and provide wider and comparable data in this area. The main direction for further research is to focus on policies that foster inequities at all levels (including organisations, communities, powiats, voivodeships, macro-regions and the nation) and elements of the built environment that are critical drivers of inequity. Furthermore, descriptive work should aim to identify priority areas for explanatory and interventional studies. --- The following are available online at https://www.mdpi.com/article/10.3390/ijerph182010856/s1, Territorial structure. --- Data Availability Statement: The data presented in this study are available on request from the corresponding author. --- Conflicts of Interest: The authors declare no conflict of interest.

The purpose of this study is to identify inequities in the distribution of socioeconomic determinants of health within Poland and their impact on the health status of Poles, as measured by mortality rate. We hypothesised that (1) there are inequities in the socioeconomic characteristics within geographically defined population groups and (2) some socioeconomic determinants of health have a particularly strong impact on the health status of Poles. Poland is administratively divided into three levels: voivodeships, powiats and gminas. We used a dataset covering all 380 powiats in Poland for the year 2018. We employed a two-stage nested Theil index and Herfindahl-Hirschman Index. In order to identify which of these determinants has the strongest impact on health, we conducted a regression analysis. The study revealed some inequities in the distribution of socioeconomic determinants of health. The mortality rate can be partly understood from variations within voivodeships in the distribution of health determinants. Important national inequalities were found in the case of two determinants, which simultaneously proved their significant impact on Poles' health in the regression analysis. Thus, type of employment and access to modern infrastructure should be of particular concern for public authorities.

Introduction Nepal is composed of different caste and ethnic groups, with unique cultures, traditions, customs, religions, beliefs, and occupations which separate the people of country from one another. Of them the Newars are the most ancient people whose antiquity goes back to the time of Gautam Buddha. They claim to be the original inhabitant of the kathmandu valley, which previously had three Newar city kingdoms, viz, Kantipur, Patan and Bhaktapur. The Gorkha prince Prithavi Narayan Shah annexed all these three kingdoms in 1825 B. S. and the valley hence forward became the capital of the new Gorkha kingdom, now called the kingdom of Nepal. The Newars comprised the majority in population in the valley. However, they have been reduced to minority. In terms of culture and religious complex, the Newar still dominate in the valley. Within the Newar community, caste system (viz, priestly castes, high castes, upper-lower castes, lower castes, unclean castes, and untouchable castes) prevails based on the division of work. Kusle is a sub-community, placed at the Newar untouchable caste hierarchy. The kusle is called also Jugi or Jogi and are said to be descended from the ascetics of Nath sect and they are also known as "Darshandhari". This caste is divided into many sections on the basis of the types of musical instruments they play. But from the point of view of marriage, ceremonies and inter-dining all of these section acts as a single endogamous --- Introduction of the Study Area Kirtipur Municipality, with an ancient and traditional life, is one of the historical places located in Kathmandu district composed of 10 different wards, where we can observe the mixed rural and urban life pattern. This Municipality is linked with Kathmandu Metropolitan City in the west and is 7 kilometre from central Kathmandu. It is one of the settlements with majority of Newar as its inhabitants and the population of other caste and ethnic group is also in significant number. The occupation of people in the rural area of the municipality is agriculture, with the remainder working in other occupations like business, foreign labour, local labour, employment, and so on classified as non-agricultural occupations. The primary crops include rice, millet, wheat, potatoes, vegetables, and so on, while the domestic animals include buffalo, cows, goats, poultry, and so on. --- Objectives The overall objective of this study is to find out the facts and figures relating to the people and their socio cultural lives of Kirtipur Municipality, whereas the specific objective of the study is to find out the social-cultural change taking place among Kusle people of Kirtipur. --- Methods The research design for this study is both exploratory and descriptive and the nature of study is qualitative and quantitative. This study is exploratory because it seeks to investigate the socio-cultural change taking place in the study area. At the same time this study is descriptive in nature since it seeks to facts and figures relating to the people and their socio cultural lives of Kirtipur Municipality. This article is prepared by using primary data, collected from Interview Schedule and secondary data were taken from previous studies, published books and other related literature. Kirtipur municipality is divided into 10 different wards. In these wards, people of several castes/ethnic groups reside. Ward no 2, 5, 7 and 10 of the Kirtipur municipality is selected for the study area. The reason for the selection of these wards is that Kusle households are not in other wards of the municipality. The actual number of Kusle household in ward no.2 is 6, in ward no. 5 is 8, in ward no. 7 is 15, and finally in ward no. 10 is 31. (Kritipur Municipality, 2078.) There are 60 households altogether in these four wards. Out of that 30 household have been included in the present study through purposive sampling method. --- Discussion and Analysis 5.1 Change in Family System: Family is a universal institution found all over the countries and societies. It is almost first institution in the human civilization. Since the primitive period of all human group it has been a small social group consisting ordinary of father, mother and one or more children. This family system is categorized into several divisions; Nuclear, Joint and Extended family system. Nuclear family consists of married couple with or without their unmarried children. Among the sampled household it was found that 16 household (53.34%) of the families were nuclear. So the researcher had found that most of the Kusle live in Nuclear. Especially Kusle of younger generation favor nuclear family system as they said that, there was no economic cooperation found in joint family and because of income disparity there was lack of 'we feeling' among the brothers. Joint family consisting of husband, wife and married children with or without their children living together in which there is sharing of joint residence, kitchen and property. 26.66% of Kusle still live in joint family. In joint family system, all the earnings are shared equally by all the family members. Due to their economic condition and population pressure, they prefer joint family system. Kusle are, however, interested in nuclear family system. An extended family consists of father, mother, and their married children living together. In other words, an extended family is that family, which consist of two or more nuclear families affiliated through an extension of the parent-child relationship rather than the husband-wife relationship. In the study area, among the 30 household, 6 of the family or 20.00% of the family are found to be an extended family which had been shown in the table 1. During the field work, it was notice that Kusle have a patrilineal society. Their descendants are traced through the line of father. Their property goes to the line of father, not mother. The son can inherit it. When a daughter gets married, she will go to her husband's house (groom's house). In the study area, according to some key informants, before 10-15 years, the Kusle people used to live in joint family. But now most of the Kusle live in nuclear family rather than joint and extended family. It is because of individualistic attitude and in the joint family they have to face many problems due to the nature of married wife, quarrels among wives of several brothers, they have to do many household works. In this connection, it can be said that due to the influence of modernization and extension of education people of the study areas prefer to live in the nuclear family. --- Change in Marriage System: Marriage, as an institution, is universal. Both man and woman are two sides of the same coin and marriage binds them together. It fulfills the physical, mental and biological needs of man and woman. Not only that but it also gives continuity to society by producing socially accepted children. Hence, marriage is a socially, economically, culturally important institution. Its types and forms differ from place to place and from one ethnic group to other in accordance with the rule and regulations and beliefs system of particular society. The following table shows the change in marriage system in the study area. The table shows that out of 30 respondents 25(83.34%) adopted arrange marriage and 5(16.66%) did love marriage in the study area. In the same way 11(36.66%) respondents give preference to arrange marriage and 19(63.34%) to love marriage. These data show that the Kusle people of Kirtipur are engaged in arranged marriage. The reasons behind engaging in arranged marriage is that in those days there is no cultural contact with outsiders and they did not want go against their parent view. They accept what their parents manage for them. But in recent periods the number of Kusle who prefer love marriage is increasing day by day. The reason behind is the education has played a crucial role. It is education which helped them to make contact with other socio-cultural life. Moreover, this social and cultural contact exerted great influence on them. --- Change in Dress and Ornaments As similar to other Newars, the Kusle people dress up in a very simple way. The Kusle wear waist-coat and Daura Suruwal, it may be cotton or woolen cloth, according to the season. The head dress of Kusle is a small skull cap of black (bhadgaunle topi or palpali topi), thinly wadded with cotton and generally turned up for an inch or so at the border and their foot wears is a simple slipper. And the women of Kusle wearing a cloth by way of petticoat, almost touching the ground, and it are covered by patasi (traditional Newari Sari, black in colour with red border). Patasi is almost 3 to 4 meters long made up in their own cottage industries. Their upper part of the body is covered by Choubandi Cholo. They, make simple makeup in their face. These are traditional dresses of Kusle people according to the respondents are suitable to their physical features and geographical environment. Since their contact with outer world, through education, social contact, media etc. their traditional patterns of dressing are gradually disappearing. Young Kusle people prefer to wear modern dresses such as pants, shirts and fashionable dress according to their economic status and interest. Likewise Kusle women have also started wearing modern saries, blouse and kurta suruwal. Young boys and girls prefer to wear fashionable clothes. Likewise, change has also come to some extent in the ornaments. Traditionally they wear simple earrings, golden ornaments, pauju in legs (made of silver). Due to the influence of outsiders the women are started wearing fashioned earrings, necklace, hand rings made of gold, on the occasion of feast, festivals and ceremonies. The wealthy women wear lots of ornaments but others wear simple ornaments. Traditionally the Kusle women didn't wear nose ring but now a days they are starting to wear nose rings too because of the contact with outsiders. And they put sindur (vermilion powder) on their head after the Bel marriage, which is quite unique feature of Kusle as well as Newari culture. Previously, Kusle did not wear pants and shirts. Even now the old Kusle people don't wear pants and shirts. --- Change in Food Habit Food is the most important thing for living. Therefore people need food to live. Every society has its own types of food habits. Likewise, Kusle have certain kinds of food habits, similar to the other Newars, Kusle are great consumer of rice buffaloes, goats, ducks and chickens. The great part of their food consists of rice but rarely do they take other types of food. They also take the mixture of green vegetables, dry vegetables, potato, and different types of beans, which is called "Quanti". They use beaten rice (chiura) for lunch. During the times of festivals and feast, they take food like chiura, raksi, different items of buff (chhyala, kachila, thalthale), different types of acchar made of potato, dry vegetables, tomato etc. the raksi is distilled from rice and wheat and families in their own houses prepare it. Kusle people of Kirtipur do not have cow and buffalo. Therefore, they do not take advantage of drinking milk and curd but now days with connection of other castes people started to drink milk and curd from collecting market. But now days those who claim themselves as a rich wealthy man stop drinking raksi, eating buff. Meat and they started leading their life differently from those of low income Kusle people. The Kusle prepare their meal themselves. Patan Pragya (Volume: 12, Number: 1, 2023) ISSN 2594-3278 96 --- Change in Religions Religion: The word called religion is extremely complex, intricate and full of paradox. It includes such facts as celebration, despair, mystery, animal sacrifice, ritual beliefs and dogmas about natural and super natural etc. Religion is a part of culture. Each and every society has its own religious tradition. Religion is an integral part of the Kusle people too. The table shows that the most of the Kusle are Buddhist and their number is 17. Similarly the 11 Kusle household also follows the Hinduism and Christian are also found in the study area. The number of Christian is 2. It doesn't mean that the Hindu Kusle don't follow Buddhist and the Buddhist Kusle don't follow Hinduism. Both follow both religions. There is religious harmony. Both visit Hindu and Buddhist religious site. Change in religion is also observed in the study area as out of 30 households 2 household change their religion. The Christian found in the sturdy area stop visit religious site and stop celebrating their traditional culture. --- Change in Religions Festivals In Nepal as in all countries, festivals are celebrated in slightly different ways, and with varying degrees of intensity depending upon locality, economic and social status education, religion and family background and personal inclination (Anderson, 1977). Nepali people celebrate different types of festivals, according to the culture and tradition. They worship different gods and goddess and they enjoy eating, drinking and feasting. The Newars celebrate all festivals, but their styles of performing a few festivals are different from other castes. Besides the three purely Hindu festivals of the Phagu, Dashain and Tihar which are observed by all classes of Nepalese people. There are certain festivals in Nepal which are peculiar to Newars, and which found no counterpart in the plains of India. (Brook: 1928,168) the Kusle celebrate Dashin in October, Tihar in November, Phagu in march, Maghe Sankranti in February, Krishna Asthami in September etc. the Kusle of Kirtipur celebrate these festivals which is similar to the other Newari castes people. --- Change in Economy of Kusle People To know the overall condition of the Kusle, we have to know the economic condition of Kusle of Kirtipur. Economic status reflects the cultural or intellectual development of certain place and it also measures the physical facilities of the people. Of course the primary (traditional) occupation of the Kusle is temple priest but now different factor are forcing certain changes in the occupational pattern. Education, modernization and consciousness itself have been the main responsible factor for the occupational mobility of the Kusle. Of the total respondents, only one Kusle (3.33%) has been working as temple priest in Bishnu Ddevi temple. Now, they have been engaged in other works like agriculture, carpenter, business, office service, driver, tailor musician etc. completely avoiding their traditional occupation. Source: Field Survey ( 2023) The table shows that, only one person (3.33%) of the total population is still continuing the traditional occupation. Rest of all changed their traditional occupation. The main occupation of the Kusle of the study area is agriculture that accounts about 43.33% of the total household. According to the obtained information, out of 30 household 27 household (90%) have their own land and the remaining 3 household (10%) do not have their own land. They cultivate land in Adhya system. Six household (20.02%) engaged in carpentry works, they work in furniture shops as a wage labour. Through which they fulfill their daily needs, which is second largest occupation adopted by Kusle in the study area. Among 30 household, 4 household (13.33%) engaged in government service in different posts. From which they manage their family's needs. Two Kusle (6.66 of the total household) engaged in business. Which includes shopkeeper of daily goods. They maintain their family from the earnings of the shop. Although the main occupation of the Kusle is agriculture and supported by carpentry, business, service, the 13.33% of the Kusle involve in other economic activities as music teacher, electrician, and driver. Due to the decreasing value of money and price of the commodity is increasing day by day. The cash becoming more necessary for fulfilling the increasing needs of the daily life. Therefore Kusle change their traditional occupation into more income generating works. However, now days, the Kusle earn sufficient money by doing different economic activities. --- Conclusion Traditional social institution of a Kusle, marriage, family, religion etc. is changing day by day. Before 10-15 years the Kusle people used to live in joint family, but at present most of the Kusle want to live in nuclear family rather than in joint and extended family. It is because of their socio-economic condition, influence of high caste people and extension of education and cultural contact with other caste people. Most of the Kusle of Kirtipur is followers of Buddhism. They celebrate feast and festivals similar to the other Newars of Kirtipur. Beside these changes, Kusle of Kirtipur is preserving their some traditional culture and in the same way they adopted some western culture simultaneously. Due to the cultural contact with other castes people as well as education, modernization and consciousness the Kusle's social as well as cultural life is changing day by day. Similarly, traditional occupation of the Kusle people was temple priest but they totally changed their occupation. The main occupation of the Kusle of Kirtipur is agriculture and some Kusle are shifting to other modern income generating activities. In the study area male are slightly economically active than female. In the same way they maintain their traditional rituals and their social prestige which they have already earned.

Change is a law of nature as nothing is static in this world. No new generation can ever be an exact replica of the old one. Social structure is never static as various internal as well as external social conditions and circumstances change the structure and function of Kusle society and culture. It can be observed that numerous research works have been carried out within Newari ethnic groups. However, the present study provides an ethnographic description and sociocultural change of the Kusle people of Kirtipur. The study employed quantitative research, with an emphasis on exploratory and descriptive research designs, and purposive sampling from four distinct wards of the municipality. The study demonstrates that due to the cultural contact with other castes people as well as education, modernization and consciousness the Kusle's social as well as cultural life is changing rapidly.

Introduction Migration due to climate change will become one of the defining trends of population movement in the 21st century (Brown, 2017). One of the most pressing global issues today revolves around water resources and the resulting migration. People often use their movement as an adaptation strategy to improve socioeconomic conditions, leading to international and internal migration movements (Nagabhatla et al., 2020b). Furthermore, the linkages between "water stress", migration, and gender represent a significant research gap, which is further exacerbated by the consequences of climate change and extreme events. Taking into account the individual Sustainable Development Goals (SDG), particularly SDG 3 (good health and well-being), SDG 6 (clean water and sanitation), SDG 7 (affordable and clean energy), SDG 10 (reduced inequalities), SDG 13 (climate action), and SDG 16 (peace, justice, and strong institutions), we have outlined recommendations and strategies while discussing multiple narratives applicable to the connections between water resources, climate and energy migration. Key points include an orientation towards long-term sustainable solutions and increasing stakeholder participation in decision-making processes. <unk>естник <unk>ара<unk>андинско<unk>о университета St. Petersburg was 25.3°C), increasing precipitation in Southeast Asia at high temperatures (above 33°C), it can be noted that these changes will have a significant impact on human health in all regions of the world, not just the poorest ones. Investor, visionary, founder of "nextbiglink", and former CEO of RAO UES Alexander Chikunov identifies two main directions of climate migration: internal, which is usually caused by natural disasters, and transnational, triggered by the intensification of risks caused by climate change (increased poverty, hunger, inequality, etc.). Natalia Makovetskaya, Director of ESG Practice at Baikal Communications Group, explains that in both cases, migration occurs involuntarily, but international law does not protect them as it does refugees. In fact, climate refugees are not legally protected (Vecherova, 2022). In the same context, Julia Kuznetsova, Associate Professor at the Faculty of Geography and Geoinformation Technology at HSE University, highlights two directions of climate migration in Russia: southern and northern. The outflow of people from the Arctic has always existed, but now it is expected to increase as glaciers melt and infrastructure in permafrost areas collapses. In the southern direction, emigration from Central Asia and the Caucasus is driven by a lack of freshwater, not only for drinking but also for irrigation. "It is projected that in the coming decades, the flow of migrants from these regions will increase. Currently, migration from these areas is driven by economic reasons, but soon climate-related reasons will be added", the expert believes (Vecherova, 2022). According to a Redfin (2021) study conducted in the United States, many residents are starting to consider climate change when determining where to live. More than 36% of respondents noted that rising sea levels were a factor in their decision to relocate, while 49% cited the need to move due to increased frequency of natural disasters. Economist Daryl Fairweather, in the same Redfin (2021) study, stated that the most competitive and expensive places to buy property will be those that are not at risk of increased natural disasters, extreme temperatures, and rising sea levels. California residents leave during the summer to avoid wildfires, while some are forced to permanently relocate due to the loss of their homes in natural disasters. Approximately 75% of residents would not consider buying a home in an area prone to various climate risks, according to Redfin (2021). As noted by the international expert community addressing this issue, there are few ideal places for comfortable human habitation on planet Earth, but even they will experience various impacts from anthropogenic factors such as waste disposal, overpopulation, lack of clean water, and as a result, a higher cost of living. However, the study of scientific works did not reveal the presence of research on the interconnection between demographic processes and the most acute global problems related to water, and as a result, climate and energy migration, as well as demoethical values as a tool for transforming sustainable development of society. Therefore, transdisciplinary, demoethical, and human-oriented research is necessary, which would emphasize the values and needs of people in managing the demographic components of the economy. The aim of the study is to: identify and analyze the causes, trends, and forecasts of climate and energy migration and propose recommendations for future research on this issue; apply demoethical values in the demographic component of the economy, which can contribute to the transition to sustainable development. --- Materials and methods The aim of this study is to explore the relationship between the demographic component of the economy (in the context of the world's most pressing water-related issues affecting climate and energy migration) and the concept of the demoethical model of sustainable development in modern society. The demoethical model focuses on organizing activities to enhance sustainable development, particularly the demoethical model relevant to Society 5.0 and Industry 5.0 organizations. By conducting a literature review and analysis, this study identifies key issues in the sustainable development of the demographic component of the economy based on the demoethical model and provides practical recommendations for stakeholders involved in digital transformation. The conceptual and methodological foundation of this research is the article "Demoethical Model of Sustainable Development of Society: A Roadmap towards Digital Transformation" (Zhanbayev et al., 2023a), which presents a new concept in the contemporary literature review and aligns with President Kassym-Jomart Tokayev's address at the meeting of the Council of Heads of States-Founders of the Interna-tional Fund for Saving the Aral Sea. The address highlighted that global climate change, the onset of a period of low water levels, and a lack of irrigation water pose a threat to the security of Central Asia. This study (Zhanbayev et al., 2023a) argues that the concept of sustainable development is based on the priority of spirituality, as the foundation of a new stage or model of societal development, which involves integrating demographic, socio-economic, and ecological components into overarching modeling processes. One example is the impact of environmental changes on individuals and their ability to adapt using alternative means of existence. The abilities and opportunities of migrants differ, meaning that some benefit while others face significant challenges. Different people migrate differently in response to drought and water shortage. Migration processes vary based on gender, age, social status, and available resources. Numerous factors such as culture, political and economic circumstances influence population migration. In recent years, empirical studies have identified population groups that either do not want or cannot engage in labor migration (Adams, 2016). The ability of these population groups to migrate may gradually decline over time, increasing the likelihood of a humanitarian crisis and the need for constant high-level intervention. The Demoethics concept can be used for environment change adaptation, creating an anti-ecological buffer, as well as risk identification and prevent the effects. --- Results --- Climate migration in Central Asia The region of Central Asia is located in the center of the Eurasian continent and includes 5 countries: the Republic of Kazakhstan, the Kyrgyz Republic, the Republic of Tajikistan, Turkmenistan, and the Republic of Uzbekistan. In Central Asia, climate change is characterized by increasing average temperatures, changing precipitation levels, and shifting extreme climate factors. The temperature increase in the region has been uneven (Table 1). The highest rates of increase in average annual air temperature are observed in Turkmenistan. In most of the territories of the Central Asian countries, the highest temperature increases occur in the spring period. The increase in extreme maximum temperatures poses a threat to agriculture, health, transportation, nature, and water usage. The shift in precipitation from summer to spring, combined with an increase in the duration of droughts, reduces water reserves, increases demand for water, and leads to water scarcity. In all five countries of Central Asia, climate change is expected to continue rising. Throughout the twentieth century, the average rate of temperature increase compared to global rates will continue to rise. By 2100, the average annual temperature may increase by 2.6%, resulting in a rise of up to 3.3°C in the summer period, according to optimistic forecasts, and by 6.8°C (up to 8.7°C in the summer period) according to pessimistic forecasts. According to international experts, the temperature in our region is increasing much faster than the global average. This is leading to a reduction in the area of glaciersthe main source of water in the Aral Sea basin. At the meeting of the heads of state of Central Asia on September 15, 2023, President Kassym-Jomart Tokayev noted that the volume of water resources in the Aral Sea has decreased by 30% over the past 50 years. By 2050, droughts in this region could cause damage equivalent to 1.3% of GDP per year, which could result in the emergence of approximately 5 million "climate" migrants in Central Asia. The region also faces problems of inefficient water use and ignoring environmental consequences. The degree of regulation of the two major rivers in our region, the Syr Darya and the Amu Darya, is extremely high. More than 80 reservoirs have been built in their basins, with a total volume of almost 65 cubic kilometers. The drying up of the Aral Sea and, as a contributing factor, the intensification of wind erosion on its drying surface, are leading to global anthropogenic causes of climate change in the Central Asian region. <unk>естник <unk>ара<unk>андинско<unk>о университета In addition to the drying up of the Aral Sea, another important specific factor affecting climate change in the region is the melting of glaciers and snow cover in the mountains, leading to desertification of the land. According to statistical data, the rate of glacier retreat in Central Asia ranges from 0.2 to 1% per year, which means that over the past 60 years, up to 30% of the Pamir and Tien Shan glaciers have melted. Thus, an increase in anthropogenic pressure is expected, along with continued extensive land use and exacerbation of various socio-ecological problems. Kassym-Jomart Tokayev noted that when formulating water policies, states should consider that water is a limited natural resource, which directly affects the wellbeing and sustainable development of the entire region. Energy migration of the EU. After the EU agreed on a price ceiling for Russian oil, the cost of Brent crude oil fell below $79 per barrel for the first time since January 2022. While Russian authorities consider possible responses to the EU's decision, energy prices for consumers, which had been steadily rising in 2022, have slightly decreased. However, since the beginning of July 2023, oil prices have started to rise again (above $90 per barrel of Brent crude), which will have a negative impact on the citizens of the UK and the EU in the upcoming winter. For example, the consortium of retailers in the UK specifies that there is a rise in prices caused by inflation due to the increase in electricity prices. The average monthly salary in the UK, according to statistical data in August 2023, is £2,260, and the average energy bill in September 2023 will be around £2,000 per year. In Germany, the average monthly salary in August 2023 is €4,330, and the average energy bill per year is €3,568. Olga Konovalova, Deputy Head of the Research and Development Department at the M.V. Lomonosov Moscow State University Center for Marine Research (Vecherova, 2022), notes that this type of migration is currently a common phenomenon and will only intensify in the future, as globalization provides individuals with diverse incomes and levels of prosperity the opportunity to migrate freely to different places at any frequency. There is a basic cost for services, and there is a cost for ecosystem services that individuals can obtain from the environment. Kuznetsov (Vecherova, 2022) notes that in this situation it is necessary to start discussing the energy transition. Thus, over the past six months (from June to November 2022), there has been an almost six-fold increase in the procurement of solar panels in European Union countries, totaling more than €2.5 billion per month. "In a way, Europe has been trying very hard to prolong its prosperity, sacrificing economic growth. But the time has come to pay for the transition to new economic models. This transition will be painful but inevitable" (Vecherova, 2022). Highlighting the economic consequences of the emerging and ongoing economic migration, statistical data shows the beginning of an economic crisis in the European Union. This is evident based on companies relocating to other countries and, including temporary relocation of citizens to warmer regions during the winter period (Vecherova, 2022). --- Discussions The correlation between Demography and the values of Demoethics In accordance with SDG 13 (climate action and peace), the African Union aims to enhance regional resilience to the negative consequences of climate change. This requires a comprehensive transformation of society and its institutions. Recent sociological surveys conducted by RPORC (Russian Public Opinion Research Center -VCIOM) and the National Energy Security Fund (Russia) indicate that the issue of climate change is becoming increasingly relevant to society. From the analysis of respondents' answers to the question "What environmental problems are currently relevant to Russia and require the most attention from the government", the Top-5 most pressing issues can be identified: water pollution, deforestation, growing landfills, air pollution, and plastic pollution (Fig. 1). According to the conducted research, 43% of respondents believe it is necessary to combat climate change "from top to bottom", ensuring appropriate legislation and control over resource usage, while only 12% of Russians believe that each individual should reduce their consumption of various resources to combat global warming. 25% of respondents stated that the recent Covid-2019 pandemic was one of the reasons why they started paying more attention to ecology, and 29% of respondents expressed the opinion that people will pay more attention to nature after the pandemic. Regarding the population's attitude towards climate change, respondents noted (Fig. 2) that climate change is partly caused by natural processes and partly by human activity -30%, climate change is mainly caused by human activity -16%, and climate change is completely caused by human activity -10%. Thus, the ROMIR study (2020) demonstrates that almost half of the respondents are convinced that both the government and individuals should make more efforts to address the climate crisis. Climate change, which leads to a deteriorating quality of life, poses additional challenges to the sociopolitical system. It is evident that the loss of family members, personal health, and housing due to climate change-related factors leads to tension in society. Climate factors (e.g., smog from natural wildfires) can exacerbate such complex political challenges for society and the state, similar to pandemics: scientists estimate that in certain cities in the US, Italy, and other countries, there is an increased mortality rate from the virus in cities with higher levels of air pollution. However, the deterioration of quality of life due to direct impact is not the only factor creating additional turbulence for any socio-political system. As global experience shows, worsening climate conditions can lead to additional migration and exacerbate social conflicts due to the rapid mixing of different subcultures and ethnic groups as a result of forced migration. The world is changing rapidly and often unpredictably. Therefore, it is necessary to note that without addressing social and environmental issues, economic sustainability cannot be achieved in any country. The interpretation of sustainable development entails ensuring social balance within society, protection of the environment, and economic efficiency in the interconnectedness of governments, business, and society, taking into account the interests of all stakeholders. One of the many systemic transformations aimed at changing the current reality is the implementation of the United Nations Sustainable Development Goals (UN SDG) (Zhanbayev et al., 2022). In order to achieve SDG 13 (climate action and its impacts), SDG 6 (provide access to water and sanitation for all), SDG 7 (ensure access to affordable, reliable, sustainable, and modern energy), SDG 3 (ensure healthy lives and promote well-being for all), and SDG 11 (make cities and human settlements inclusive, safe, resilient, and sustainable), measures are proposed for developing the demography component. Migration is an objective process that should be managed using government regulatory mechanisms, actively utilizing the tools of Demoethics and digitization in relation to water, climate, and energy policies of the state. Migration is an objective process that should be managed with the help of state regulation mechanisms, actively using the tools of demoethics and digitalization in relation to water, climate, and energy policies of the state. Climate change is caused by human activity and threatens life on Earth as we know it. Further inevitable deterioration of the environment will lead to increased migration flows from Central Asian countries. In the ranking of vulnerability to climate change compiled by the World Bank, Tajikistan ranks first among all countries in Eastern Europe and Central Asia, Kyrgyzstan is in third place, Uzbekistan is in sixth place, and Turkmenistan is in seventh place. The fight against climate change should be carried out, first of all, by the state, creating a legislative framework (Zhanbayev et al., 2023b) based on the concept of Demoethics values and ensuring control over the use of natural resources. Unfortunately, the formation of public demand for environmental modernization faces a number of problems caused, among other things, by the imperfection of state climate and environmental policy. We consider Demoethical Communication Management in the context of implementing measures to prevent worst-case climate scenarios and introducing new climate policies as the most effective tool at the present time. Thus, the values of Demoethics contribute to the adaptation of society to negative consequences, being a tool of stabilization in the conditions of the new climate reality, which is currently being formed and affects all countries existing today. The practical implementation of the research is provided by Demoethics, as a new model for the transformation of society in the context of the global climate crisis. The proposed developed paradigms of Demoethics values, such as demography, democracy and demoeconomics, can play an important role in the implementation of the Sustainable Development Goals and, as a result, ensure the quality of life and competitiveness of the population. This study examines the issues of the economic componentdemography, using Demoethic values such as "spirituality and morality", "responsibility", "justice", "rationality",and "security" to create favorable conditions for improving demographic indicators, which are presented in Table 2. One of the key concepts of the Demoethic approach is the value ("spirituality and morality", "responsibility", "justice", "rationality", and "security"). In our research, we have determined that the key organization responsible for the well-being of the economy and society is responsible for creating and implementing values in sustainable development. --- Values of Demoethics Characteristics of Demoethics values Spiritual-moral The influence of the "spiritual and moral" aspect on the demographic indicators of society can be very great in various circumstances: People's ethical and moral beliefs can influence how they decide important life issues: getting married, having children, and raising children. In those societies where marriage and family are valued, spiritual and moral values help preserve the family in difficult moments, and also, as a rule, set people up to give birth and raise as many children as possible. Thus, spiritual and moral values contribute to an increase in the birth rate and an increase in the population. Spiritual and moral attitudes can also influence the decision to terminate a pregnancy (abortion). In countries with strong religious traditions, where abortion is prohibited due to national or religious beliefs, the number of abortions is much lower compared to societies where these traditions are less significant. Moral values and spiritual culture may still influence the aging population and overall life expectancy. In countries where caring for elderly parents is valued, young people are more likely to take on this responsibility, and accordingly, life expectancy there increases. Responsibility "Demographic responsibility" plays an important role in the sustainable development of society under the following circumstances: Socio-economic sustainability: When a society has a stable economic and social environment, people feel more confident about the future and are more likely to be demographically responsible. They are much more likely to plan and have children, not just one, but several, as well as take care of their health, provide decent upbringing and education. Access to Resources and Services: When governments and business communities provide their citizens with quality access to social services such as health care, education, housing and a standard of living, people are more likely to engage in responsible parenting and parenting. Having the support and necessary resources helps them meet the needs of their family and create favorable conditions for giving birth and raising their heirs. Social values and norms: Cultural and social factors greatly influence demographic responsibility. In some societies, the state may encourage demographic responsibility through government programs to provide social support to families with children or through measures to stimulate the birth rate. In addition, public opinion and norms also play a role in shaping people's behavior regarding having and raising children. Examples: In some developed countries, such as Sweden and Norway, high levels of demographic responsibility are associated with wide access to social services, gender equality and low levels of poverty. These societies actively support families with children by providing a wide range of child benefits, parental leave and child care. Other countries, such as Japan and Germany, are facing declining birth rates and an aging population. In such circumstances, governments implement measures to stimulate the birth rate to cope with demographic challenges, such as providing maternity capital, preferential conditions and support for kindergartens. This is an example of demographic responsibility on the part of the state. --- Justice Demographic equity is an important aspect of social justice and is often the focus of government policies and programs at the national level to build a more equal and just society. Several demographic trends have emerged during the Covid-19 pandemic that can be considered in the context of demographic justice: Fertility: Many countries have seen a decline in their birth rate during the pandemic. This may be explained by stress, economic uncertainties and restrictions, and the immediate effects of the disease on pregnant women. This aspect affects demographic justice, since a decrease in the birth rate can lead to increased demographic inequality and an increase in the problem of population aging. Mortality: Covid-19 caused more serious consequences in older people and people with certain medical conditions than in younger and healthy people. This has led to a significant increase in mortality in these population groups. Demographic justice requires ensuring access to health care, support and protection for vulnerable groups. Migration: The pandemic has also affected international and domestic migration flows. The introduction of restrictions on the movement of people and the closure of borders led to a decrease in migration. At the same time, vulnerable groups such as refugees and migrants found it more difficult to access health care and social protection. This contributed to the violation of demographic justice. Overall, demographic equity has been at risk during the Covid-19 pandemic due to declining fertility rates, increased mortality among vulnerable groups, and restrictions on migration. Ensuring equal opportunities and protection for all groups of the population has been an important aspect of demographic justice in the field of demographic processes during the Covid-19 pandemic. <unk>естник <unk>ара<unk>андинско<unk>о университета --- Values of Demoethics Characteristics of Demoethics values Rationality Demographic rationalities are a concept associated with the rationale and making of decisions that are based on the demographic characteristics and characteristics of a particular population group. Various social, economic and cultural factors can influence such rational decisions. Demographic rationalities in society include the following aspects: Fertility and family policy: Currently, in many developed countries there is a decline in fertility and an aging population, so it is necessary to apply rational and thoughtful solutions that will be aimed at supporting families and stimulating fertility (for example, subsidies for children's goods and services, benefits for young families, increase in maternity leave). Western European countries such as Sweden and Norway have legislated family support policies, including parental leave for both parents, subsidies for child care and other measures. Japan and Germany are developing and implementing programs aimed at increasing fertility, such as financial incentives for young families and support for women who do not want to continue their careers. The general goals of demographic rationality include understanding the needs and requirements of different demographic groups in order to develop appropriate policies and programs. Population aging: In societies where the ratio of the elderly population to the working population is high, there is a demographic rationality in decision-making aimed at improving the living conditions of older people. For example, the development of infrastructure, social protection and medical care for the elderly population. Youthful population: In countries with young populations, decisions related to education, health care and job creation for young people will be demographically rational. For example, developing the education system, ensuring access to quality medical care and promoting youth entrepreneurship. Migration: In societies with high levels of international migration, rational decisions may be related to the development of migration policies and the integration of migrants into society. For example, creating support and adaptation programs for new migrants, as well as taking into account the importance of intercultural dialogue and promoting the inclusion of migrants into society. Many countries attract foreign workers to fill labor shortages in certain industries. Thus, in the countries of Western Europe and North America in recent decades, programs have been introduced to help attract skilled workers from other countries, which has led to an increase in the number of migrants. This can have both positive and negative impacts on fertility and family policies in receiving and sending countries. Conflicts and humanitarian crises arise in various regions of the world, resulting in people being forced to leave their home countries in search of security and stability. A country's acceptance of refugees may affect demographics due to changes in population structure and cultural differences. --- Security The "security" aspect can influence the demographic indicators of society as follows: Military Conflicts: Military actions typically result in deaths, displacement, and destruction of social and economic infrastructure. This, in turn, provokes a decrease in the birth rate and an increase in mortality, which affects the demographic indicators of society. Crime and Violence: High levels of crime and violence can have a negative impact on the demographics of a society. Due to rising crime rates, people are forced to leave their homes and countries. This leads to increased migration and negatively affects the birth rate and mortality rate. Disadvantage and Poverty: Disadvantage and poverty often prevent people from accessing basic health services, education and social programs. This leads to an increase in diseases, poor health, high mortality and low birth rates, which negatively affects the demographic indicators of society. Examples: During the Syrian civil war, many people were forced to leave their homes and country due to hostilities. This led to an increase in migration and negatively affected the demographic indicators of Syria. In some cities with high crime rates, such as Caracas in Venezuela, violence can lead to high death rates and decreased birth rates. This negatively affects the demographic situation in the city. --- Note -(developed by the authors): the proposed preliminary results of applying the values of Demo-ethics to the demographic component of the economy will be improved within the framework of the study --- Conclusions Technological transformation is one of the key solutions for managing economic, environmental, and political risks. The current situation with climate change, as well as the plans of other countries, require a radical transformation of Kazakhstan's socio-economic development and the establishment of new relationships with foreign partners based on dialogue on climate issues and technological modernization. The current situation seems to have deeply undermined people's trust and faith in the data economy, its practices, and players (Rantanen, 2019). This lack of trust and its consequencesinformation exchange limitations and data falsificationthreaten data collection and, therefore, the entire foundation of the data economy (Punj, 2019). Thus, a new human-centered and transparent approach to the data economy is necessary to overcome the lack of trust and risks to an ethical society (Koskinen et al., 2017). Therefore, our re-search raises questions about Demotic values and a human-centered model of sustainable development, demotic management of the demographic component of the economy in the context of climate and energy migrations. In addition, the "Low Carbon Transformation of the Economy", based on the values of demo-ethics, will allow to avoid economic crises and political risks in the future society or minimize them, and most importantly, prevent already manifested unethical consequences by respecting the rights of citizens, including the right to a favorable environment. Following the strategic goals and based on the values of Demoethics, actions should be directed towards creating favorable conditions in the most vulnerable sectors of the economy: 1) Water resources 2) Energy 3) Agriculture and land use 4) Healthcare 5) Natural ecosystems and biodiversity 6) Reducing the risks of "climate-dependent natural disasters". Within the framework of achieving the Sustainable Development Goals (SDG), the country, civil society, and international institutions have taken on commitments for the period until 2030 to fulfill a whole range of tasksfrom reducing extreme poverty to ensuring environmental sustainability and universal education, reducing child mortality, and promoting gender equality. Despite the fact that adaptation activities directly contribute to the achievement of SDG 13, aimed at taking measures to combat climate change, it should be considered as part of a strategy to consolidate the achievements of countries in implementing all other SDG based on the value of Demoethics. We expect that the results of this study can be useful in supporting targeted policies and the transformation of demographic components of the economy. --- <unk>дат<unk>а: <unk>ерттеуд<unk> ма<unk>сат<unk> -<unk>о<unk>амн<unk> т<unk>ра<unk>т<unk> даму<unk>н <unk>ерту <unk>рал<unk> рет<unk>нде су ресурстар<unk>мен, климатт<unk>, <unk>нер<unk>етикал<unk> к<unk>-<unk>онмен <unk>не демо<unk>тикал<unk> <unk>нд<unk>л<unk>тармен <unk>а<unk>лан<unk>ст<unk> <unk>лемде<unk> е<unk> <unk>тк<unk>р м<unk>селелер контекс<unk>нде<unk> демо<unk>ра<unk>и<unk>л<unk> <unk>ро<unk>естерд<unk> <unk>ара <unk>а<unk>лан<unk>с<unk>н <unk>ерттеу. <unk>д<unk>с<unk>: <unk>а<unk>алада демо<unk>ра<unk>и<unk>л<unk>, <unk>леуметт<unk>к-<unk>кономикал<unk> <unk>не <unk>коло<unk>и<unk>л<unk> ком<unk>оненттерд<unk> <unk>ал<unk> <unk>ел<unk>к модел<unk>деу <unk>ро<unk>естер<unk>не <unk>р<unk>кт<unk>руден т<unk>рат<unk>н <unk>о<unk>амн<unk> т<unk>ра<unk>т<unk> даму<unk>н<unk> <unk>а<unk>а ке<unk>е<unk>н<unk> немесе <unk>а<unk>а модел<unk>н<unk> не<unk> рет<unk>нде ру<unk>ани<unk>тт<unk> <unk>ас<unk>мд<unk>на не<unk>дел<unk>ен демо<unk>тика т<unk>р<unk>мдамас<unk> <unk>олдан<unk>лд<unk>. <unk>ор<unk>т<unk>нд<unk>: <unk>с<unk>н<unk>л<unk>ан т<unk>с<unk>л т<unk>ра<unk>т<unk> дамуд<unk> <unk>а<unk>сарту <unk>н<unk>нде<unk> <unk>метт<unk> <unk>мдаст<unk>ру<unk>а <unk>а<unk>ттал<unk>ан. <unk>та<unk> а<unk>т<unk>анда, <unk>о<unk>ам 5.0 мен <unk>ндустри<unk> 5.0 <unk>мдар<unk> <unk>н ма<unk>д<unk> <unk>ол<unk> та<unk>лат<unk>н демоетикал<unk> модел<unk> <unk>с<unk>н<unk>л<unk>ан. <unk>де<unk>иеттерд<unk> <unk>олу <unk>не талдау ар<unk>л<unk> <unk>л <unk>ерттеу <unk>тикал<unk> <unk>л<unk> не<unk>нде <unk>кономикан<unk> демо<unk>ра<unk>и<unk>л<unk> <unk>рамдас <unk>л<unk>н<unk> т<unk>ра<unk>т<unk> даму<unk>н<unk> не<unk> м<unk>селелер<unk>н ан<unk>та<unk>д<unk> <unk>не <unk>и<unk>рл<unk> транс<unk>орма<unk>и<unk>а <unk>ат<unk>сат<unk>н м<unk>ддел<unk> тара<unk>тар <unk>н <unk>рактикал<unk> <unk>с<unk>н<unk>стар <unk>еред<unk>. <unk>р<unk>мдама: <unk>лд<unk> т<unk>ра<unk>т<unk> даму салас<unk>нда<unk> ма<unk>саттар<unk>а (<unk>) <unk>ол <unk>етк<unk>у <unk>е<unk>ер<unk>нде а<unk>аматт<unk> <unk>о<unk>ам мен <unk>ал<unk>арал<unk> институттар 2030 <unk>л<unk>а де<unk>н<unk> ке<unk>е<unk>е аса кеде<unk>л<unk>к к<unk>рсетк<unk>тер<unk>н <unk>с<unk>артудан <unk>аста<unk> <unk>коло<unk>и<unk>л<unk> т<unk>ра<unk>т<unk>л<unk> <unk>ен <unk>ал<unk>а <unk>рде<unk> <unk>л<unk>м <unk>еруд<unk> <unk>амтамас<unk> ету<unk>е, <unk>алалар <unk>л<unk>м<unk>н <unk>с<unk>арту<unk>а <unk>не <unk>ендерл<unk>к те<unk>д<unk>кт<unk> <unk>амтамас<unk> ету<unk>е де<unk>н<unk> м<unk>ндеттерд<unk> т<unk>тас с<unk>ектр<unk>н ор<unk>ндау <unk>о<unk>н<unk>а <unk>не м<unk>ндеттемелер <unk>а<unk>лдад<unk>. <unk>е<unk>мделу <unk>мет<unk> климатт<unk> <unk>еру<unk>не <unk>арс<unk> к<unk>рес <unk>аралар<unk>н <unk>а<unk>лдау<unk>а <unk>а<unk>ттал<unk>ан <unk> 13-ке <unk>ол <unk>етк<unk>у<unk>е т<unk>келе<unk> <unk>ал етет<unk>н<unk>не <unk>арамастан, он<unk> демо<unk>тика <unk>нд<unk>л<unk>на не<unk>дел<unk>ен <unk>арл<unk> <unk>ас<unk>а <unk>-н<unk> <unk>е<unk>е ас<unk>руда елдерд<unk> <unk>ет<unk>ст<unk>ктер<unk>н <unk>ек<unk>ту страте<unk>и<unk>с<unk>н<unk> <unk>л<unk> рет<unk>нде <unk>араст<unk>ру <unk>а<unk>ет. <unk>т<unk>лет<unk>н н<unk>ти<unk>елер: <unk>с<unk> <unk>ерттеуд<unk> н<unk>ти<unk>елер<unk> ма<unk>сатт<unk> са<unk>сатт<unk> <unk>олдау <unk>не <unk>кономикан<unk> демо<unk>ра<unk>и<unk>л<unk> ком<unk>оненттер<unk>н транс<unk>орма<unk>и<unk>лау <unk>н не<unk> <unk>ола алад<unk>. --- <unk>ннота<unk>и<unk>: <unk>ел<unk>: <unk>ел<unk> исследовани<unk> <unk>акл<unk>аетс<unk> в и<unk>у<unk>ении в<unk>аимосв<unk>и демо<unk>ра<unk>и<unk>ески<unk> <unk>ро<unk>ессов в контексте наи<unk>олее остр<unk> миров<unk> <unk>ро<unk>лем, св<unk>анн<unk> с водн<unk>ми ресурсами, с климати<unk>еско<unk>, <unk>нер<unk>ети<unk>е

Object: To study the interrelation of demographic processes in the context of the most pressing global issues related to water resources, climate, energy migration, and demoethical values as a tool for the transformation of sustainable development in society. Methods: The study utilizes the concept of Demoethics, which is based on the priority of spirituality as the basis for a new stage or model of sustainable development in society, which involves integrating the demographic, socioeconomic, and ecological components into the overall processes of modeling. Findings: The proposed approach focuses on organizing activities to improve sustainable development. In particular, a demographic model is presented, which is important for the Society 5.0 and Industry 5.0 organizations. Through literature review and analysis, this research identifies key issues in the sustainable development of the demographic component of the economy based on the demoethical model and provides practical recommendations for stakeholders involved in digital transformation. Conclusions: In order to achieve the Sustainable Development Goals (SDG) for countries, civil society, and international institutions by 2030, commitments have been made to accomplish a range of tasks, from reducing extreme poverty to ensuring environmental sustainability and universal education, reducing child mortality, and promoting gender equality. Despite the fact that adaptation activities directly contribute to achieving SDG 13, which focuses on taking action against climate change, it should be seen as part of a strategy to strengthen a country's achievements in implementing all other SDGs based on the value of Demoethics. Expected results: The results of this study can serve as a basis for supporting targeted policies and the transformation of demographic components of the economy.

н<unk>ти<unk>елер: <unk>с<unk> <unk>ерттеуд<unk> н<unk>ти<unk>елер<unk> ма<unk>сатт<unk> са<unk>сатт<unk> <unk>олдау <unk>не <unk>кономикан<unk> демо<unk>ра<unk>и<unk>л<unk> ком<unk>оненттер<unk>н транс<unk>орма<unk>и<unk>лау <unk>н не<unk> <unk>ола алад<unk>. --- <unk>ннота<unk>и<unk>: <unk>ел<unk>: <unk>ел<unk> исследовани<unk> <unk>акл<unk>аетс<unk> в и<unk>у<unk>ении в<unk>аимосв<unk>и демо<unk>ра<unk>и<unk>ески<unk> <unk>ро<unk>ессов в контексте наи<unk>олее остр<unk> миров<unk> <unk>ро<unk>лем, св<unk>анн<unk> с водн<unk>ми ресурсами, с климати<unk>еско<unk>, <unk>нер<unk>ети<unk>еско<unk> ми<unk>ра<unk>ие<unk>, и демо<unk>ти<unk>ески<unk> <unk>енносте<unk> как инструмент транс<unk>орма<unk>ии усто<unk>иво<unk>о ра<unk>вити<unk> о<unk>ества. <unk>етод<unk>: <unk> ра<unk>оте ис<unk>ол<unk>ована <unk>он<unk>е<unk>и<unk> демо<unk>тики, <unk>а<unk>иру<unk>а<unk>с<unk> на <unk>риоритете ду<unk>овности, как основ<unk> ново<unk>о <unk>та<unk>а или ново<unk> модели усто<unk>иво<unk>о ра<unk>вити<unk> о<unk>ества, котора<unk> <unk>акл<unk>аетс<unk> в инте<unk>ра<unk>ии демо<unk>ра<unk>и<unk>еско<unk>, со<unk>иал<unk>но-<unk>кономи<unk>еско<unk> и <unk>коло<unk>и<unk>еско<unk> составл<unk>и<unk> в о<unk>есистемн<unk>е <unk>ро<unk>есс<unk> моделировани<unk>. <unk>е<unk>ул<unk>тат<unk>: <unk>редла<unk>аем<unk> <unk>од<unk>од <unk>окусируетс<unk> на ор<unk>ани<unk>а<unk>ии де<unk>тел<unk>ности <unk>о улу<unk>ени<unk> усто<unk>иво<unk>о ра<unk>вити<unk>. <unk> <unk>астности, <unk>редставлена демо<unk>ти<unk>еска<unk> модел<unk>, ва<unk>на<unk> дл<unk> ор<unk>ани<unk>а<unk>и<unk> <unk>ества 5.0 и <unk>ндустрии 5.0. <unk>осредством о<unk>ора и анали<unk>а литератур<unk> данное исследование о<unk>редел<unk>ет кл<unk>ев<unk>е во<unk>рос<unk> усто<unk>иво<unk>о ра<unk>вити<unk> демо<unk>ра<unk>и<unk>еско<unk>о ком<unk>онента <unk>кономики на основе де<unk>ти<unk>еско<unk> модели и <unk>редоставл<unk>ет <unk>ракти<unk>еские рекоменда<unk>ии дл<unk> <unk>аинтересованн<unk> сторон, у<unk>аству<unk>и<unk> в <unk>и<unk>рово<unk> транс<unk>орма<unk>ии. <unk>вод<unk>: <unk> рамка<unk> дости<unk>ени<unk> <unk>еле<unk> в о<unk>ласти усто<unk>иво<unk>о ра<unk>вити<unk> (<unk>) стран<unk> <unk>ра<unk>данское о<unk>ество и ме<unk>дународн<unk>е институт<unk> на <unk>ериод до 2030 <unk>ода <unk>рин<unk>ли на се<unk> о<unk>ател<unk>ства <unk>о в<unk>олнени<unk> <unk>ело<unk>о с<unk>ектра <unk>ада<unk> -от сокра<unk>ени<unk> <unk>ока<unk>ателе<unk> кра<unk>не<unk> <unk>едности до о<unk>ес<unk>е<unk>ени<unk> <unk>коло<unk>и<unk>еско<unk> усто<unk>ивости и всео<unk>е<unk>о о<unk>ра<unk>овани<unk>, сокра<unk>ени<unk> детско<unk> смертности и о<unk>ес<unk>е<unk>ени<unk> <unk>ендерно<unk>о равенства. <unk>есмотр<unk> на то, <unk>то ада<unk>та<unk>ионна<unk> де<unk>тел<unk>ност<unk> на<unk>р<unk>му<unk> с<unk>осо<unk>ствует дости<unk>ени<unk> <unk>-13, на<unk>равленно<unk> на <unk>рин<unk>тие мер <unk>о <unk>ор<unk>е с и<unk>менением климата, ее ну<unk>но рассматриват<unk> как <unk>аст<unk> страте<unk>ии <unk>о <unk>акре<unk>лени<unk> дости<unk>ени<unk> стран в осу<unk>ествление все<unk> дру<unk>и<unk> <unk> на основе <unk>енности демо<unk>тики. <unk>естник <unk>ара<unk>андинско<unk>о университета <unk>идаем<unk>е ре<unk>ул<unk>тат<unk>: <unk>е<unk>ул<unk>тат<unk> данно<unk>о исследовани<unk> мо<unk>ут <unk>ослу<unk>ит<unk> осново<unk> дл<unk> <unk>оддер<unk>ки <unk>елена<unk>равленно<unk> <unk>олитики и транс<unk>орма<unk>ии демо<unk>ра<unk>и<unk>ески<unk> ком<unk>онентов <unk>кономики. <unk>л<unk>ев<unk>е слова: <unk>ели усто<unk>иво<unk>о ра<unk>вити<unk> (<unk>), усто<unk>ивое ра<unk>витие, демо<unk>ти<unk>еские <unk>енности, климати<unk>еска<unk> ми<unk>ра<unk>и<unk>, <unk>нер<unk>ети<unk>еска<unk> ми<unk>ра<unk>и<unk>, <unk>е<unk>о<unk>асност<unk> <unk>ентрал<unk>но<unk> <unk>ии, водн<unk>е ресурс<unk>, ка<unk>ество <unk>и<unk>ни, транс<unk>орма<unk>и<unk>.

Object: To study the interrelation of demographic processes in the context of the most pressing global issues related to water resources, climate, energy migration, and demoethical values as a tool for the transformation of sustainable development in society. Methods: The study utilizes the concept of Demoethics, which is based on the priority of spirituality as the basis for a new stage or model of sustainable development in society, which involves integrating the demographic, socioeconomic, and ecological components into the overall processes of modeling. Findings: The proposed approach focuses on organizing activities to improve sustainable development. In particular, a demographic model is presented, which is important for the Society 5.0 and Industry 5.0 organizations. Through literature review and analysis, this research identifies key issues in the sustainable development of the demographic component of the economy based on the demoethical model and provides practical recommendations for stakeholders involved in digital transformation. Conclusions: In order to achieve the Sustainable Development Goals (SDG) for countries, civil society, and international institutions by 2030, commitments have been made to accomplish a range of tasks, from reducing extreme poverty to ensuring environmental sustainability and universal education, reducing child mortality, and promoting gender equality. Despite the fact that adaptation activities directly contribute to achieving SDG 13, which focuses on taking action against climate change, it should be seen as part of a strategy to strengthen a country's achievements in implementing all other SDGs based on the value of Demoethics. Expected results: The results of this study can serve as a basis for supporting targeted policies and the transformation of demographic components of the economy.

Introduction Understanding interactions among sets of entities often represented as complex networks, is a central research task in many data-intensive scientific fields, including Statis-tics, Machine learning, Social sciences, Biology, Psychology, and Economics (Watts and Strogatz, 1998;Barabási and Albert, 1999;Albert and Barabási, 2002;Jackson et al., 2008;Girvan and Newman, 2002;Shmulevich et al., 2002;Bickel and Chen, 2009;Bullmore and Sporns, 2009;Rubinov and Sporns, 2010;Carrington et al., 2005;Borgatti et al., 2009;Wasserman et al., 1994;Lazer, 2011). However, a majority of methodological and applied studies have only considered interactions of one type among a set of entities of the same type. More recent studies have pointed to the heterogeneous and multimodal nature of such interactions, whereby a complex networked system is composed of multiple types of interactions among entities that themselves are of multiple types (Kivelä et al., 2014;Boccaletti et al., 2014;Mucha et al., 2010;Paul and Chen, 2016;Paul et al., 2020b;Sengupta and Chen, 2015;Sun et al., 2009;Liu et al., 2014;Ferrara et al., 2014;He et al., 2014;Nickel et al., 2016;Paul et al., 2020a). Social relationships are known to affect individual behaviors and outcomes including dementia (Fratiglioni et al., 2000), decision making (Kim and Srivastava, 2007), adolescent smoking Mercken et al. (2010), and online behavior choices (Kwon et al., 2014). At the same time, individual attributes, such as race, age, and gender, can affect whether and how people form friendships or romantic partnerships (Dean et al., 2017;McPherson et al., 2001). The effect of social relationships on individual behaviors is observed through disparities in the outcomes across different individuals when their friendship structures differ. The reciprocal is also observed through disparities in the friendship structures when individuals' attributes differ. Therefore, flexible joint modeling of social relationships and individual behaviors and attributes is needed to investigate their interrelationships effectively. A number of popular models for social networks have been extended to incorporate nodal covariates in the literature, including, exponential random graph models (ERGMs) (Lusher et al., 2013), stochastic blockmodels (Mele et al., 2019;Sweet, 2015) and latent space models (Fosdick and Hoff, 2015;Krivitsky and Handcock, 2008;Hoff et al., 2002;Austin et al., 2013). In these models, the social network links are treated as dependent variables, and the effects of nodal covariates on the probability of network ties are subsequently estimated. Alternatively, social influence models use the node-level attributes as the dependent variables and estimate the effects of the social network on the attributes (Dorans and Drasgow, 1978;Leenders, 2002;Robins et al., 2001;Sweet and Adhikari, 2020;Frank et al., 2004;Fujimoto et al., 2013;Shalizi and Thomas, 2011;VanderWeele, 2011;VanderWeele and An, 2013;Bramoullé et al., 2009;Goldsmith-Pinkham and Imbens, 2013;Shalizi and Thomas, 2011). A different approach is to develop a joint modeling framework where different types of data are integrated by jointly modeling them as the dependent variables. In network science, the joint modeling framework has been proposed to model multi-view or multiplex networks, where multiple types of relations are observed on the same set of nodes, using stochastic block models (Barbillon et al., 2015;Kéfi et al., 2016), and latent space models (Gollini and Murphy, 2016;Arroyo et al., 2019;Salter-Townshend and McCormick, 2017;D'Angelo et al., 2018;Zhang et al., 2020). In these models, latent variables are used to explain the probability of a node being connected with other nodes in multiple types of relations. When dependencies can be assumed across different layers, the common node representation is a flexible framework to summarize multiple types of information. In many cases, high-dimensional multivariate covariates with complex latent structures are available in addition to a connected network. In particular, often a social network is observed along with individuals' attributes or behavioral outcomes. In this case, two types of relations, the social network relations and various types of individual attributes, are observed among two types of nodes, the person nodes, and the attribute nodes. To jointly model such data, a multivariate normal distribution was fitted by Fosdick and Hoff (2015) to the latent variables from the social network and the observed covariates. This work is in spirit the closest to our proposed model. However, it restricts the covariates to be normally distributed, and it does not take into account the multiple latent dimensions of the covariates. A dynamic version of Fosdick and Hoff (2015) can be found in Guhaniyogi et al. (2020) with possibilities to accommodate both categorical and continuous attributes. The most important distinction of our model from this line of work is that we use a second set of latent variables, the latent attribute variables, in addition to the latent person variables to summarize the information associated with each attribute. This modeling framework allows for the joint latent space, where two types of nodes are interactive instead of one. Other related works that jointly model heterogeneous networks are recently seen with stochastic block models including Huang et al. (2020); Sengupta and Chen (2015). In this paper, we propose a joint latent space model for heterogeneous and multimodal networks with multiple types of relations among multiple types of nodes. The proposed Attribute Person Latent Space Model (APLSM) merges information from the social network and the multivariate covariates by assuming that the probabilities of a node being connected with other same-type and different-type nodes are explained by latent variables. This model has a wide range of applications. For example in computer science, it is of interest to summarize relational data, e.g. likes and followers in social media with other user information such as personalities, health outcomes, online behavior choices, etc. In economics and business, it is of interest to summarize consumer informa-tion with their geographic networks and social networks. We demonstrate APLSM with a data set on the French financial elites (Kadushin, 1995) available to download from http://moreno.ss.uci.edu/data.html#ffe. To fit the APLSM, we propose a Variational Bayesian Expectation-Maximization (VBEM) algorithm Blei et al. (2017). As an intermediate step we also develop a VBEM algorithm for fitting latent space models to bipartite networks. The variational methods enable the models to be fitted to large networks with high dimensional attributes, while our simulations show the accuracy of the methods. The remainder of this paper is organized as follows. In section 2, we introduce latent space models for bipartite networks and develop the variational inference approach to estimate the model. In section 3, we introduce the joint latent space model for the social network and the multivariate covariates and extend the variational inference method to the joint model. In section 4, we assess the performance of the estimators with a simulation study, and in section 5, we apply the proposed methodology to the French financial elite data. Finally, in section 6, we summarize our findings. --- Latent Space Models Development of Latent Space Models (LSM) for social networks can be traced back to Hoff et al. (2002)'s latent distance and latent projection models. Both models assume that nodes can be positioned in a D-dimensional latent space and that the probability of an edge between two nodes depends on their closeness. The closer the two nodes, the less likely they form a connection. Conditional on these latent positions, the probability of two nodes forming an edge is independent of all the other edges in the network. In the latent distance model, the Euclidean distances are used to describe the relationships between the nodes; whereas, in the latent projection model, scaled vector products are used to describe the relationships between nodes. Let Y I denote the N <unk> N adjacency matrix of the social network among N individuals. The (i, j) th element of the matrix Y I, denoted as y I i j is 1 if person i and person j are related, for i, j = <unk>1, 2,..., N <unk> and i <unk> j. Let U be a N <unk> D matrix of latent person position, each row of which is a D dimensional vector u i = (u i1, u i2,..., u iD ) indicating the latent position of person i in the Euclidean space. The latent distance model for a binary social network Y I can be written as: Y I i j |(U, <unk> 0 ) <unk> Bernoulli(g(<unk> i j )), g(<unk> i j ) = exp(<unk> 0 -|u i -u j | 2 ) 1 + exp(<unk> 0 -|u i -u j | 2 ). We assume u i to be the latent person variable, u i iid <unk> N (0, <unk> 2 0 I D ), where <unk> 0, <unk> 2 0 are unknown parameters that need to be estimated and I D is the D dimensional identity matrix. The probability of an edge increases as the Euclidean distance between the two nodes decreases. Here, we use the squared Euclidean distances |u i -u j | 2 instead of the Euclidean distances following (Gollini and Murphy, 2016). It has been shown in Gollini and Murphy (2016) that squared Euclidean distances are computationally more efficient and that the latent positions obtained using squared Euclidean distances are extremely similar to those obtained using Euclidean distances. Variations of the latent space model are further developed in Hoff (2005Hoff (, 2008Hoff (, 2009Hoff (, 2018) ) and Ma et al. (2020). --- Latent Space Model for Bipartite Networks Data on the attributes of the individuals in the network can be seen as a bipartite network with directed edges between two types of nodes. The development of the latent space model for bipartite networks includes a bipartite version of the latent cluster random effects model in the latentnet package (Krivitsky and Handcock, 2008). In addition, the latent space model for a dynamic bipartite network was introduced by (Friel et al., 2016) to study the interlocking directorates in Irish companies. In the rest of this section, we introduce a Variational Bayesian EM algorithm for fitting a latent space model to binary bipartite networks (BLSM). Let Y I A denote the N <unk> M bipartite network, whose (i, a) th element y I A ia is 1 if person i has attribute a, for i = <unk>1, 2,..., N <unk> and a = <unk>1, 2,..., M <unk>. Let V be a M<unk>D matrix of latent attribute positions, each row of which is a D dimensional vector v a = (v a1, v a2,..., v aD ) indicating the latent position of attribute a in the Euclidean space. The latent distance model for the bipartite network Y I A can be written as: Y I A ia |(U, V, <unk> 0 ) <unk> Bernoulli(g(<unk> ia )), g(<unk> ia ) = exp(<unk> 1 -|u i -v a | 2 ) 1 + exp(<unk> 1 -|u i -v a | 2 ), We assume u i iid <unk> N (0, <unk> 2 0 I D ), v a iid <unk> N (0, <unk> 2 1 I D ), and <unk> 1, <unk> 0 and <unk> 1 to be unknown parameters. The parameter <unk> 1 accounts for the density of the bipartite network. The probability of a positive response increases as the Euclidean distance between the attribute node and the person node decreases. --- Variational Bayesian EM for the Bipartite Network We are interested in the posterior inference of the latent variables u i and v a following the distance model conditioning on the observed bipartite network. The (conditional) posterior distribution is the ratio of the joint distribution of the observed data and unobserved latent variables to the observed data likelihood P(U, V |Y I A ) = P(Y I A |U, V)P(U, V) P(Y I A ). We can characterize the distribution of latent positions and thus obtain the point and interval estimates by computing this posterior distribution. The variational inference algorithm is commonly used to estimate latent variables whose posterior distribution is intractable (Beal et al., 2003;Attias, 1999;Beal et al., 2006;Blei et al., 2017). In network analysis, the variational approach has been proposed for the stochastic blockmodel (Daudin et al., 2008;Celisse et al., 2012), the mixed-membership stochastic blockmodel (Airoldi et al., 2008), the multi-layer stochastic blockmodel (Xu et al., 2014;Paul and Chen, 2016), the dynamic stochastic blockmodel (Matias and Miele, 2016), the latent position cluster model (Salter-Townshend and Murphy, 2013) and the multiple network latent space model (Gollini and Murphy, 2016). Here, we propose a Variational Bayesian Expectation Maximization (VBEM) algorithm to approximate the posterior of the person and the attribute latent positions using the bipartite network. We propose a class of suitable variational posterior distributions for the conditional distribution of (U, V |Y I A ) and obtain a distribution from the class that minimizes the Kulback Leibler (KL) divergence from the true but intractable posterior. We assign the following variational posterior distributions: q(u i ) = N ( <unk>i, <unk>0 ) and q(v a ) = N (<unk> a, <unk>1 ) and set the joint distribution as q(U, V |Y I A ) = N i=1 q(u i ) M a=1 q(v a ), where <unk>i, <unk>0, <unk>a, <unk>1 are the parameters of the variational distribution, known as variational parameters. We can estimate the variational parameters by minimizing the Kullback-Leiber (KL) divergence between the variational posterior q(U, V |Y I A ) and the true posterior f (U, V |Y I A ). Minimizing the KL divergence is equivalent to maximizing the following Evidence Lower Bound (ELBO) function Blei et al. (2017) = - N <unk> i=1 KL[q(u i )| f (u i )] - M <unk> a=1 KL[q(v a )| f (v a )] + E q(U,V,<unk> 1 |Y I A ) [log f (Y I A |U, V, <unk> 1 )] = - 1 2 DN log(<unk> 2 0 ) -N log(det( <unk>0 )) - N tr( <unk>0 ) 2<unk> 2 0 - N i=1 <unk>i T <unk>i 2<unk> 2 0 - 1 2 D M log(<unk> 2 1 ) -M log(det( <unk>1 )) - M tr( <unk>1 ) 2<unk> 2 1 - M a=1 <unk>a T <unk>a 2<unk> 2 1 + 1 2 (M D + N D) + E q(U,V |Y I A ) [log f (Y I A |U, V)].(1) After applying Jensen's inequality (Jensen, 1906), a lower-bound on E q(U,V |Y I A ) [log f (Y I A |U, V)] is given by, E q(U,V |Y I A ) [log f (Y I A |U, V, <unk> 1 )] <unk> N <unk> i=1 M <unk> a=1 y ia <unk>1 -tr( <unk>0 ) -tr( <unk>1 ) -( <unk>i -<unk>a ) T ( <unk>i -<unk>a ) - N <unk> i=1 M <unk> a=1 log 1 + exp( <unk>1 ) det(I + 2 <unk>0 + 2 <unk>1 ) 1 2 exp -( <unk>i -<unk>a ) T (I + 2 <unk>0 + 2 <unk>1 ) -1 ( <unk>i -<unk>a ). We use the Variational Expectation-Maximization algorithm (Jordan et al., 1999;Baum et al., 1970;Dempster et al., 1977) to maximize the ELBO function. Following the variational EM algorithm, we replace the E step of the celebrated EM algorithm, where we compute the expectation of the complete likelihood with respect to the conditional distribution f (U, V |Y I A ), with a VE step, where we compute the expectation with respect to the best variational distribution (obtained by optimizing the ELBO function) at that iteration. The detailed procedures are as follows. We start with the initial parameter, <unk> (0) = <unk>(0) 1, and <unk>(0) i, <unk>(0) 0, <unk>(0) a, <unk>(0) 1, and then we iterate the following VE (Variational expectation) and M (maximization) steps. During the VE step, we maximize the ELBO(q(U), q(V), <unk>) with respect to the variational parameters <unk>i, <unk>a, <unk>0 and <unk>1 given the other model parameters and obtain ELBO(q * (U), q * (V), <unk>). During the M step, we fix <unk>i, <unk>a, <unk>0 and <unk>1 and maximize the ELBO(q(U), q(V), <unk>) with respect to <unk>1. To do this, we differentiate ELBO(q(U), q(V), <unk>) with respect to each variational parameter. We obtain closed form update rules by setting the partial derivatives to zero while introducing the first-and second-order Taylor series expansion approximation of the log functions in ELBO(q(U), q(V), <unk>) (see detailed derivations in supplementary material). The Taylor series expansions are commonly used in the variational approaches. For example, three first-order Taylor expansions were used by Salter-Townshend and Murphy (2013) to simplify the Euclidean distance in the latent position cluster model, and first-and second-order Taylor expansions were used by Gollini and Murphy (2016) to simplify the squared Euclidean distance in LSM. Following the previous publications using Taylor expansions, we approximate the three log functions in our ELBO(q(U), q(V), <unk>) function to find closed form update rules for the variational parameters. Define the function F ia = N <unk> i=1 M <unk> a=1 log 1 + exp( <unk>1 ) det(I + 2 <unk>0 + 2 <unk>1 ) 1 2 exp -( <unk>i -<unk>a ) T (I + 2 <unk>0 + 2 <unk>1 ) -1 ( <unk>i -<unk>a ). The closed form update rules of the (t + 1)th iteration are as follows t+1), <unk>(t+1) 0 and <unk>(t+1) VE-step: Estimate <unk>i (t+1), <unk>a( 1 by minimizing ELBO(q(U), q(V), <unk>) <unk>(t+1) i = 1 2<unk> 2 0 + M <unk> a=1 y ia ) I + 1 2 H ia ( <unk>(t) i ) -1 M <unk> a=1 y ia <unk>(t) a + 1 2 H ia ( <unk>(t) i ) <unk>(t) i - 1 2 G ia ( <unk>(t) i ) <unk>(t+1) a = 1 2<unk> 2 1 + N <unk> i=1 y ia I - 1 2 H ia (<unk> (t) a ) -1 N <unk> i=1 y ia <unk>(t) i - 1 2 G ia (<unk> (t) a ) <unk>(t+1) 0 = N 2 N 2 1 <unk> 2 0 + N <unk> i=1 M <unk> a=1 y ia I + G ia ( <unk>(t) 0 ) -1 <unk>(t+1) 1 = M 2 M 2 1 <unk> 2 1 + N <unk> i=1 M <unk> a=1 y ia I + G ia ( <unk>(t) 1 ) -1,(2) where G I A ( <unk>i (t) )and G I A (<unk> a (t) ) are the partial derivatives (gradients) of F I A with respect to <unk>i and <unk>a, evaluated at <unk>i (t) and <unk>a (t), respectively. In G I A ( <unk>i (t) ), the subscript I A indicates that the gradient is of function F I A, and the subscript i in <unk>i (t) indicates that the gradient is with respect to <unk>i, evaluated at <unk>i (t). Similarly, H I A ( <unk>i (t) ) and H I A (<unk> a (t) ) are the second-order partial derivatives of F I A with respect to <unk>i and <unk>a, evaluated at <unk>i (t) and <unk>a (t), respectively. M-step: Estimate <unk>(t+1) 1 with the following closed form solution, <unk>(t+1) 1 = N i=1 M a=1 y I A ia -g I A ( <unk>(t) 1 ) + <unk>(t) 1 h I A ( <unk>(t) 1 ) h I A ( <unk>(t) 1 ),(3) where g I A ( <unk>(t) 1 ) is the partial derivative (gradient) of F I A with respect to <unk>1, evaluated at <unk>(t) 1 ; and h I A ( <unk>(t) 1 ) is the second-order partial derivative of F I A with respect to <unk>1, evaluated at <unk>(t) 1. 3 Joint Modeling of Social Network and Multivariate --- Attributes Next we define the attributes and persons latent space model (APLSM), which uses a single joint latent space to combine and summarize the information contained in both the person-person social network Y I, and person-attribute bipartite network formed by the high dimensional covariates Y I A. We assume that the persons and attributes can be positioned in an attribute-person joint latent space, which is a subset of the D dimensional Euclidean space R D. The latent person positions U is a shared latent variable that simultaneously affects both the social network and the multivariate covariates. In APLSM, we extend the conditional independence assumption of LSM by assuming that the probability of two nodes forming a connection in both matrices Y I and Y I A is independent of all other connections given the latent positions of the two nodes involved. In the APLSM, the joint distribution of the elements of the social network and the multivariate covariates can be written as p(Y I, Y I A |U, V, <unk> 0, <unk> 1 ) = N i=1 N j=1, j<unk>i p 1 (y I i, j |<unk> I i, j ) N i=1 M a=1 p 2 (y I A i,a |<unk> I A i,a ), E (y I i, j |<unk> I i, j ) = g 1 (<unk> I i, j ), E (y I A i,a |<unk> I A i,a ) = g 2 (<unk> I A i,a ), where g i (•) are the link functions, and p i (•|•) are the parametric families of distributions suitable for the type of data in the matrices. We set the priors u i iid <unk> N (0, <unk> 2 0 I D ), and v a iid <unk> N (0, <unk> 2 1 I D ). <unk> 0, <unk> 1, <unk> 2 0, <unk> 2 1 are unknown parameters. Different levels of <unk> 0 and <unk> 1 account for different densities of the two networks. <unk> I i, j = <unk> 0 -|u i -u j | 2, and <unk> I A i,a = <unk> 1 -|u i -v a | 2,(4) In Equation 4, squared euclidean distance forms are assumed for <unk> I i, j and <unk> I A i,a. If the data in Y I and Y I A are binary, then the link functions g 1 (<unk> I i, j ) and g 2 (<unk> I A i,a ) are logistic inverse link functions, i.e., g 1 (<unk> I i, j ) = While it is common for the edges in the friendship networks to be binary, the multivariate covariates can be more general. If the data in Y I A are of discrete numerical scales, they can be modeled with other parametric families. For example, we can use g 2 (<unk> I A i,a ) = exp(<unk> I A i,a ) as the Poisson inverse link function to model count data in Y I A, and thus p 2 (y I A i,a |<unk> I A i,a ) becomes the PDF of the Poisson distribution. If the data in Y I A are normally distributed, then the link function is the identity link. Alternatively, we can model the presence (or absence) of an edge separately from the weight of the edge (if it is present). For example, a zero inflated normal distribution was used by Sewell and Chen (2016) to model weighted edges, and the same goal was achieved by Agarwal and Xue (2020) using a combination of Bernoulli distribution and a non-parametric weight distribution. In a similar fashion, the APLSM can be used to handle weighted edges. A zero inflated Poisson model for the distribution of y I A i,a |<unk> I A i,a can be seen as follows: p 3 (y I A i,a |<unk> I A i,a ) = (1 -(<unk>(<unk> I A i,a )) (y I A i,a =0) <unk> (<unk>(<unk> I A i,a ) exp(-<unk>(<unk> I A i,a ))<unk>(<unk> I A i,a ) y I A i,a y I A! i,a <unk>(<unk> I A i,a ) = exp(<unk> I A i,a ) 1 + exp(<unk> I A i,a ) <unk>(<unk> I A i,a ) = exp(<unk> I A i,a ). --- The Variational Bayesian EM algorithm The VBEM algorithm for fitting APLSM can be derived following steps similar to section 2.2. We start by noting that the posterior distribution of APLSM is as follows: P(U, V |Y I, Y I A ) = P(Y I, Y I A |U, V)P(U, V) P(Y I, Y I A ). We assign the following variational posterior distribution: q(u i ) = N ( <unk>i, <unk>0 ) and q(v a ) = N (<unk> a, <unk>1 ). We set the joint distribution as q(U, V |Y I, Y I A ) = N i=1 q(u i ) M a=1 q(v a ), where <unk>i, <unk>0, <unk>a, <unk>1 are the parameters of the distribution, known as variational parameters. The Evidence Lower Bound (ELBO) function for APLSM is (see detailed derivations in the Supplementary Materials) ELBO = -E q(U,V,<unk> 0,<unk> 1 |Y I,Y I A ) [ log q(U, V, <unk> 0, <unk> 1 |Y I, Y I A ) log p(U, V, Y I, Y I A |<unk> 0, <unk> 1 ) ] = - <unk> q(U, V, <unk> 0, <unk> 1 |Y I, Y I A ) log q(U, V, <unk> 0, <unk> 1 |Y I, Y I A ) f (U, V, <unk> 0, <unk> 1 |Y I, Y I A ) d (U, V, <unk> 0, <unk> 1 ) = - <unk> N i=1 q(u i ) M a=1 q(v a ) log N i=1 q(u i ) M a=1 q(v a ) f (Y I, Y I A |U, V, <unk> 0, <unk> 1 ) N i=1 f (u i ) M a=1 f (v a ) d (U, V, <unk> 0, <unk> 1 ) = - N <unk> i=1 <unk> q(u i ) log q(u i ) f (u i ) du i - M <unk> a=1 <unk> q(v a ) log q(v a ) f (v a ) dv a + <unk> q(U, V, <unk> 0, <unk> 1 |Y I, Y I A ) log f (Y I, Y I A |U, V, <unk> 0, <unk> 1 )d (U, V, <unk> 0, <unk> 1 ) = - N <unk> i=1 KL[q(u i )| f (u i )] - M <unk> a=1 KL[q(v a )| f (v a )] + E q(U,V,<unk> 0,<unk> 1 |Y I,Y I A ) [log f (Y I, Y I A |U, V, <unk> 0, <unk> 1 )] = - 1 2 DN log(<unk> 2 0 ) -N log(det( <unk>0 )) - N tr( <unk>0 ) 2<unk> 2 0 - N i=1 <unk>i T <unk>i 2<unk> 2 0 - 1 2 D M log(<unk> 2 1 ) -M log(det( <unk>1 )) - M tr( <unk>1 ) 2<unk> 2 1 - M a=1 <unk>a T <unk>a 2<unk> 2 1 + 1 2 (M D + N D) + E q(U,V |Y I,Y I A ) [log f (Y I, Y I A |U, V)].(5) After applying Jensen's inequality (Jensen, 1906), a lower-bound on the third term is given by, E q(U,V |Y I,Y I A ) [log f (Y I, Y I A |U, V, <unk> 0, <unk> 1 )] <unk> N <unk> i=1 M <unk> a=1 y ia <unk>1 -tr( <unk>0 ) -tr( <unk>1 ) -( <unk>i -<unk>a ) T ( <unk>i -<unk>a ) + N <unk> i=1 N <unk> j=1, j<unk>i y i j <unk>0 -2 tr( <unk>0 ) -( <unk>i -<unk> j ) T ( <unk>i -<unk> j ) - N <unk> i=1 M <unk> a=1 log 1 + exp( <unk>1 ) det(I + 2 <unk>0 + 2 <unk>1

In many application problems in social, behavioral, and economic sciences, researchers often have data on a social network among a group of individuals along with high dimensional multivariate measurements for each individual. To analyze such networked data structures, we propose a joint Attribute and Person Latent Space Model (APLSM) that summarizes information from the social network and the multiple attribute measurements in a person-attribute joint latent space. We develop a Variational Bayesian Expectation-Maximization estimation algorithm to estimate the posterior distribution of the attribute and person locations in the joint latent space. This methodology allows for effective integration, informative visualization, and prediction of social networks and high dimensional attribute measurements. Using APLSM, we explore the inner workings of the French financial elites based on their social networks and their career, political views, and social status. We observe a division in the social circles of the French elites in accordance with the differences in their individual characteristics.

I A |U, V)].(5) After applying Jensen's inequality (Jensen, 1906), a lower-bound on the third term is given by, E q(U,V |Y I,Y I A ) [log f (Y I, Y I A |U, V, <unk> 0, <unk> 1 )] <unk> N <unk> i=1 M <unk> a=1 y ia <unk>1 -tr( <unk>0 ) -tr( <unk>1 ) -( <unk>i -<unk>a ) T ( <unk>i -<unk>a ) + N <unk> i=1 N <unk> j=1, j<unk>i y i j <unk>0 -2 tr( <unk>0 ) -( <unk>i -<unk> j ) T ( <unk>i -<unk> j ) - N <unk> i=1 M <unk> a=1 log 1 + exp( <unk>1 ) det(I + 2 <unk>0 + 2 <unk>1 ) 1 2 exp -( <unk>i -<unk>a ) T (I + 2 <unk>0 + 2 <unk>1 ) -1 ( <unk>i -<unk>a ) - N <unk> i=1 N <unk> j=1, j<unk>i log 1 + exp( <unk>0 ) det(I + 4 <unk>0 ) 1/2 exp -( <unk>i -<unk> j ) T (I + 4 <unk>0 ) -1 ( <unk>i -<unk> j ).(6) In addition to F ia, we also take into account F i F i = N <unk> i=1 N <unk> j=1, j<unk>i log 1 + exp( <unk>0 ) det(I + 4 <unk>0 ) 1/2 exp -( <unk>i -<unk> j ) T (I + 4 <unk>0 ) -1 ( <unk>i -<unk> j ). The closed form update rules of the (t + 1)th iteration are as follows t+1), <unk>(t+1) 0 and <unk>(t+1) Compute estimates <unk>0 and <unk>1 using Equations 11 and 12 VE-step: Estimate <unk>i (t+1), <unk>a( 1 by minimizing ELBO(q(U), q(V), <unk>) <unk>(t+1) i = 1 2<unk> 2 0 + N <unk> j=1, j<unk>i (y i j + y ji ) + M <unk> a=1 y ia ) I + H i ( <unk>(t) i ) + 1 2 H ia ( <unk>(t) i ) -1 (7) N <unk> j=1, j<unk>i (y i j + y ji ) <unk> j + M <unk> a=1 y ia <unk>(t) a -G i ( <unk>(t) i ) + H i ( <unk>(t) i ) + 1 2 H ia ( <unk>(t) i ) <unk>(t) i - 1 2 G ia ( <unk>(t) i ) (8) <unk>(t+1) a = 1 2<unk> 2 1 + N <unk> i=1 y ia I - 1 2 H ia (<unk> (t) a ) -1 N <unk> i=1 y ia <unk>(t) i - 1 2 G ia (<unk> (t) a ) <unk>(t+1) 0 = N 2 N 2 1 <unk> 2 0 + 2 N <unk> i=1 N <unk> j=1 y i j + N <unk> i=1 M <unk> a=1 y ia I + G i ( <unk>(t) 0 ) + G ia ( <unk>(t) 0 ) -1 (9) 3: Compute estimates <unk>0 and <unk>1 using Equations 9 and 10 4: Compute estimates <unk>i and <unk>a using Equations 7 and 8 5: Compute estimates KL using Equations 5 and 6 6: KL dis <unk> KL t /KL t-1 7: t <unk> t + 1 8: end while 9: return [ <unk>0, <unk>1,, <unk>i, <unk>0, <unk>a, <unk>1, KL] <unk>(t+1) 1 = M 2 M 2 1 <unk> 2 1 + N <unk> i=1 M <unk> a=1 y ia I + G ia ( <unk>(t) 1 ) -1,(10) where G I ( <unk>i (t) ), G I A ( <unk>i (t) )and G I A (<unk> a (t) ) are the partial derivatives (gradients) of F I, F I A and F I A with respect to <unk>i, <unk>i and <unk>a, evaluated at <unk>i (t), <unk>i (t) and <unk>a (t), respectively. In G I ( <unk>i (t) ), the subscript I indicates that the gradient is of function F I, and the subscript i in <unk>i (t) indicates that the gradient is with respect to <unk>i, evaluated at <unk>i (t). Similarly, H I ( <unk>i (t) ), H I A ( <unk>i (t) ) and H I A (<unk> a (t) ) are the second-order partial derivatives of F I, F I A and F I A with respect to <unk>i, <unk>a, <unk>i and <unk>a, evaluated at <unk>i (t), <unk>a (t), <unk>i (t) and <unk>a (t), respectively. M-step: Estimate <unk>(t+1) 0 and <unk>(t+1) 1 with the following update rules, <unk>(t+1) 0 = N i=1 N j=1 y I i j -g I ( <unk>(t) 0 ) + <unk>(t) 0 h I ( <unk>(t) 0 ) h I ( <unk>(t) 0 ) (11) <unk>(t+1) 1 = N i=1 M a=1 y I A ia -g I A ( <unk>(t) 1 ) + <unk>(t) 1 h I A ( <unk>(t) 1 ) h I A ( <unk>(t) 1 ), (12 ) where g I ( <unk>(t) 0 ) and g I A ( <unk>(t) 1 ) are the partial derivatives (gradients) of F I and F I A with respect to <unk>0 and <unk>1, evaluated at <unk>(t) 0, <unk>(t) 1 ; and h I ( <unk>(t) 0 ) and h I A ( <unk>(t) 1 ) are the second-order partial derivatives of F I and F I A with respect to <unk>0 and <unk>1, evaluated at <unk>(t) 0, <unk>(t) 1. Figure 1: The distributions of the AAEs when <unk> 0 = 2 and <unk> 1 = 1.5. --- Simulation Study In this section, we conduct a simulation study to evaluate the proposed VBEM algorithm's performance for the APLSM. We compare APLSM with two baseline approaches, namely LSM using variational inference proposed by Gollini and Murphy (2016), and LSM for bipartite networks (BLSM) using variational inference developed in Section 2.2. We expect that with the inclusion of the information in Y I A, the APLSM will exhibit a stronger model recovery fit than the LSM for the social network. Similarly, we expect the APLSM to be more successful in model recovery than the BLSM due to the addition of information from Y I. To assess whether the proposed algorithm can recover the true link probabilities, we use the estimated latent positions to the pairwise distances based on the true latent positions. Even though the true latent positions cannot be exactly recovered in any latent space model due to unidentifiability, we expect a successful algorithm to be able to recover the true distances Sewell and Chen (2015). If the estimated latent positions preserve the nodes' relative positions in the latent space, we expect the proportion of the estimated pairwise distances to the true pairwise distances to be close to 1. The design of the simulation is as follows. We first generate data following Equation 4. To do this, we set <unk> 0 and <unk> 1 to be 1 and the number of attributes and persons to be 50. We sample Z I and Z A from the multivariate normal distributions using the above parameter values. We produce true link probabilities between persons and between attributes and persons using different sets of <unk> values. Different <unk> values are associated with different densities of the data. Then, we generate Y I and Y I A matrices. Each entry of the matrices is independently generated from the Bernoulli distribution using the corresponding link probability. We apply the APLSM with the VBEM estimator to the generated data and obtain the latent positions' posterior distributions and estimates for the fixed parameters. We use the posterior means as the point estimates of the latent positions to obtain the estimated probabilities. We also fit LSM to Y I and BLSM to Y I A. Using the posterior means as point estimates of the latent positions, we obtain the estimated probabilities for Y I and Y I A, respectively. We repeat this process 200 times and report the results. when <unk> 0 = -1 and <unk> 1 = 0.5. In Figure 1, we present the boxplots over 200 simulations of averages of the absolute differences (AAEs) across entries in Y I and Y I A for APLSM, LSM and BLSM when <unk> 0 = 2 and <unk> 1 = 1.5. We see that the AAEs are generally smaller when an APLSM is fitted to the data compared to AAEs when an LSM is fitted. Therefore there is a strong improvement in model recovery using APLSM for Y I compared with using LSM. Clearly, this improvement results from the added attribute information that consolidates the latent person position estimates. Similarly, the AAEs are smaller based on fitting the APLSM than AAEs based on fitting the BLSM. Again, this suggests a big improvement in model recovery using APLSM for Y I A than using BLSM. In this case, the added social network consolidates the latent person position estimates resulting in a better model recovery for APLSM. In Figure 2, we present the distributions of the differences between the estimated <unk> values and the true <unk> values. The true <unk> values are.5 and 0 for <unk> 0 and <unk> 1. The distribution of the differences for each <unk> is centered around 0, indicating little bias in the <unk> estimates. Both also are relatively narrow, implying that the estimated <unk> values are precise and close to the true <unk> values. Finally, in Figure 3, we compare the pairwise distances based on the estimated latent positions with those based on the true latent positions. The distance between nodes i and j using estimated <unk>i and <unk> j should be close to the true distance between nodes i and j using u i and u j if the VBEM estimation algorithm successfully maintains and recovers the relationship between nodes i and j. As shown in both plots in Figure 3, the distributions of the ratios of the estimated and true pairwise distances are narrow and centered around 1, implying satisfactory recoveries of the nodes' relationships with each other through the estimated latent positions. --- Application We apply APLSM to the French financial elite dataset collected by Kadushin and de Quillacqi to study the friendship network among top financial elites in France during the last years of the Socialist government Kadushin (1995). We first introduce the data and then describe our results in detail. --- Data The data were collected through interviews for people who held leading positions in major financial institutions and frequently appeared in financial section press reports. The friendship information was collected by asking the interviewees to name their friends in the social context. Kadushin and de Quillacqi then identified an inner circle of 28 elites from the initial sample based on their influence and their perceived eliteness by other participants. The resulting friendship network is a symmetric adjacency matrix. The data also contains additional background information, including age, a complex set of post-secondary education experiences, place of birth, political and cabinet careers, political party preference, religion, current residence, and club memberships. Two aspects of the elites' "prestige" include whether the person is named in the social register and whether the person has a particle ("de") in front of either his (no woman was in the inner circle), his wife's, his mother's, or his children's names. Having "de" in the name is associated with nobility. Father's occupation is one of the variables used to reflect an elite's social class. Fathers' occupation is considered "high" if the father is in higher management, a professional, an industrialist, or an investor. Unfortunately, upon communications with the original author, we found that the coding procedures regarding some variables have been lost, including Finance Ministry information, religion, etc. We end up with 13 binary variables including information on education ("Science Po", "Polytechniqu", "University" and "Ecole Nationale d'Administration"), career ("Inspection General de Finance" and "Cabinet"), class ("Social Register", "Father Status", "Particule"), politics ("Socialist", "Capitalist" and "Centrist") and "Age" after excluding the lost or the unrelated information, i.e., mason and location, which are not associated with the social network based on Kadushin (1995) (location is not considered to be related to the social network after adjusting for multiple comparisons). "Age" was converted into a binary variable following Kadushin (1995), where a group of elites was considered of older age with an average birth year of 1938. We will use ENA as an abbreviation for Ecole Nationale d'Administration. The respected in France as it leads to membership in the ENA, where the grands corps, which are the French civil service elites, including the Inspection General de Finance, etc-recruit its members (Kadushin, 1995). --- Previous Approaches The authors in Kadushin (1995) first used multidimensional scaling to draw the friendship network's sociogram. Then they applied Quadratic Assignment Procedure regressions and correlations to test each background variable's association with the social network. Based on the social network, two clusters were identified, which the authors called the left and the right moieties. The dependence between the social network and background information was understood through comparisons of the elites between the left and the right moiety. The elites in the right moiety were found to have a higher social class (upper-class parentage with high social standing), to be older (average birth year of 1929), and to have fewer appointments in public offices. The left moiety elites were more likely to be ENA graduates, grand corps members, cabinet members, treasury service members, socialists, and younger (average birth year of 1938). Using the APLSM, we will construct a joint latent space which will allow us to jointly model elites' friendship connections and their background information. Using the APLSM, we will also replicate Kadushin (1995)'s left and the right moiety, adding simultaneous interpretation for the division in the elite circle. Furthermore, we observe an additional division within the left moiety using APLSM, which provides opportunities for new hypotheses. --- Analyses and Findings We fit the APLSM model to the social network Y I and the covariates Y I A of the French financial elites dataset. For comparison, we also fit the social network Y I and the covariates Y I A separately using LSM. We fit the latent space model to the friendship network using the variational inference proposed in Gollini and Murphy (2016). For the multivariate covariates, we fit the bipartite latent space model (BLSM) using the variational inference method we developed in Section 2.2. In Figure 4, we present the estimated latent person and attribute positions, <unk>i and <unk>a from fitting the BLSM (left) and from fitting the APLSM (right). In Figure 5, we exclusively compare the resulting latent person positions using only the social network, Y I (left), only the multivariate covariates, Y I A (middle), and both the social network and the multivariate covariates, Y I and Y I A (right). An angular rotation of the latent friendship space is applied to match the latent person positions in the joint latent space. A congruence coefficient of 0.96 was found between the two sets of latent positions. Used as a measure of dimension similarity, a congruence coefficient of 0.95 and above indicates that the dimensions are identical. A congruence coefficient of 0.91 was found between the latent person positions using the BLSM and the latent person positions using the APLSM. As expected, the latent person positions' structure using APLSM more closely resembles the structure using LSM than that using BLSM. --- Assess Model Fit To assess the model fit of APLSM to the data, we obtain the area under the receiver operating characteristic curve (AUC) of predicting the presence or absence of a link from the estimated link probabilities. The receiver operating characteristic curves (ROCs) and the AUC values for Y I and Y I A are presented in Figure 6. The results show satisfactory fit for both matrices. In addition, we assess whether the APLSM captures the structure of the social network and the dependencies among the attributes and the persons' attributes information in general. In Figure 7, we assess whether the APLSM can capture the friendship structure in the social network. In the left panel, we compare the elite's distances to the center with the sum totals of their friendship connections. We plot the rankings of the distances to center against the rankings of the total friendship connections. In this way, we assess whether the APLSM preserves the elites' social hierarchy without being distracted by the distribution differences between Euclidean distances and friendship counts. A solid reference line is drawn to illustrate the relationship between the two. The intercept for the solid line equals to N, the highest ranking, and the slope equals to -1. As the solid line roughly goes through the center of the scatter plot, we know that the rankings of the distances to the center mirror the rankings of the friendship counts in the opposite direction. The Spearman's rank correlation between the Euclidean distances and friendship counts is -0.53. In the right panel of Figure 7, we present the box plot of elites' pairwise differences in the absence of friendship versus when they are friends. We can see that the pairwise distances are generally smaller between elites who are friends with a median of 0.75, compared to the pairwise distances between elites who are not friends with a median of 1.60. This observation shows that pairwise distances between elites using the APLSM distinguish between whether the elites are friends or not. In Figure 8, we assess the APLSM's ability to capture the dependencies among attributes. In the left panel, we plot the rankings of attributes' distances to center against the rankings of the attribute sum scores. As can be seen, the two types of rankings roughly follow the solid line with an intercept of M, the highest ranking, and a slope of -1, suggesting a mirroring of the two rankings in the opposite direction. The Spearman's rank correlation between the attribute's distances to center and attribute sum scores is -0.92. This observation suggests that attributes' distances to the center capture the overall response rates of the attributes. In the right panel of Figure 8, we plot the rankings of the pairwise distances against the rankings of the attribute correlation values. As can be seen, the two types of rankings center around the dashed line with an intercept of 80, the highest ranking, and a slope of -1, suggesting a mirroring of the two rankings in the opposite direction; the Spearman's rank correlation between the two is -0.62. This result shows that the pairwise distances between attributes using the APLSM capture the correlations between the attributes. Using Figure 9 (left), we assess whether the APLSM is able to differentiate the presence of an attribute from the absence. We compare the pairwise distances between attributes and persons when the attribute is present to the pairwise distances between persons and attributes when the attribute is absent. We note that when the attributes are present, the distances are generally smaller with a median of 1.13, than those when the attributes are absent with a median of 1.71. Therefore the pairwise distances between persons and attributes using the APLSM distinguish the present from the absence of attributes. --- Redefine the Left and Right Moiety To replicate the left and right moiety in Kadushin (1995), we apply K-means clustering (Hartigan et al., 1979) In Figure 10, we show the latent person and attribute positions colored by the resulting two clusters. The solid lines represent the observed friendship connections between pairs of elites. We can see that a more densely connected network can be found on the left side than the right. Figure 9 (right) shows the probability of a pair of elites being friends in each cluster and between the two clusters. The median probability of friendship connections in the left cluster is roughly 0.29; the median probability in the right cluster is roughly 0.22; while between the two clusters, the median probability is roughly 0.05. This result shows that there is slightly stronger connectivity among elites in the left cluster than in the right cluster. Also, there is a separation of French elites based on social connectivity into the two clusters. In Figure 10, we also see that attributes: Socialists, Inspector General de France, ENA, Cabinet, University, Polytechnique and Center (Centrists) are found in the left (red) cluster, suggesting that the elites in this cluster are more likely to obtain higher education, hold top office positions and identify as socialists or centrists. On the other hand, attributes such as Right (Capitalist), Father Status, Social Register, Particule, Science Po, and Age are found in the right (green) cluster, which suggest that the elites in the right cluster are more likely to be of higher class (being present in the social register, having particule in their names and fathers of higher status), older age and identify as politically right. Two clusters based on APLSM largely correspond to the left and right moieties found in Kadushin (1995). However, in contrast to Kadushin (1995), the two clusters here are joint clusters with both attributes and persons. The addition of the attributes provides more meaning to the division of french elites. We see the division of the social circles based on party affiliations, class, education, and career. This finding was also observed in the previous study though Kadushin (1995) compared the background information for elites between the left and right moiety after the analysis of the network (The 12.1% of the variance in only four attribute variables was explained through regression). Because we are able to use relevant attribute information in the clustering process, there is more confidence in our observation of the division. Besides, the approximate credible intervals of the person and attribute nodes using the APLSM is much smaller than those using the BLSM and the LSM. The drivers behind the division automatically and systematically emerge from the data. --- Career, Politics and Class The APLSM captures how social relations between French elites and their career, politics, and class information are related. In the previous section, we have replicated the division in the French financial elites between the left and right moiety. Though not mentioned in the previous study Kadushin (1995), we believe that the presence of a third cluster might be justified given the visible separation between the ENA attribute and the Polytechniqu attribute in the left moiety and our prior knowledge about the French education system. Therefore, we run the k-means clustering algorithm with k = 3. The algorithm is again implemented with the kmeans function in the kknn P package (Hechenbichler and Schliep, 2004) with 100 random starting positions. Approximately 63.9% of the total variance is explained by the three clusters, with 13.5% of the variance explained by the added cluster. As before, the resulting three clusters contain both attributes and persons. As we know, Polytechnique, an alternative for the Science Po, indicates an alternative career for the French financial elites. We display the estimated probabilities of the presence of an attribute by elites in each of the three clusters in Figure 12. For the ENA attribute, we observe clear distinctions among the three clusters, with the highest probabilities found for elites in the ENA cluster and the lowest probabilities found for elites in the HighClass cluster. For party affiliations (Socialist, Centrist, and Capitalist), we observe higher probabilities of socialists being in the ENA cluster than the Polytechnique cluster and extremely low probabilities of socialists being in the HighClass cluster. We observe slightly higher probabilities of capitalists being in the HighClass cluster than the ENA cluster, though both probabilities are much higher than capitalists being in the Polytechnique cluster. We also observe much higher probabilities of centrists being Polytechnique graduates than being ENA graduates or from a high-class background. For class background and being part of the social register, we observe lower probabilities of being in the social register for elites in the ENA and the Polytechnique clusters than for elites in the HighClass cluster. These additional information can be used to generate new hypotheses for further understanding of the French elite class. In Figure 13, we display the probability of a pair of elites being friends with a person between ENA graduates and non-ENA graduates as the ENA graduates center around the ENA attribute. --- Discussions and conclusion The APLSM outlined in this article constitutes a principle strategy for jointly analyzing social networks and high dimensional multivariate covariates. We have argued for and presented evidence that a joint analysis of friendships and individual outcomes is crucial in understanding human behaviors. In particular, using APLSM, we have analyzed the French financial elite data, replicated the division in the elite circle between the left and right moiety, explained and consolidated the division using career, class, and political information. Furthermore, we identified an additional cluster in the joint latent space. Therefore, we believe that APLSM can significantly help researchers understand how friendships and multivariate covariates are intertwined and inspire further model development in this area. Supplementary Material for "Joint Latent Space Model for Social Networks with Multivariate Attributes". 7 The Estimation Procedure for APLSM --- Derivation of KL Divergence We set the variational parameter as <unk> = <unk>0, <unk>1 and <unk>i, <unk>0, <unk>a, <unk>1, where q(u i ) = N ( <unk>i, <unk>0 ), and q(v a ) = N (<unk> a, <unk>1 ). We set the variational posterior as: q(U, V |Y I, Y I A ) = N i=1 q(u i ) M a=1 q(v a ) The Kullback-Leiber divergence between the variational posterior and the true posterior is: where each of the components are calculated as follows: KL[q(U, V, <unk> 0, <unk> 1 |Y I, Y I A )| f (U, V, <unk> 0, <unk> 1 |Y I, Y I A )] = <unk> q(U, N <unk> i=1 KL[q(u i )|| f (u i )] = - N <unk> i=1 <unk> q(u i ) log f (u i ) q(u i ) du i = - N <unk> i=1 <unk> q(u i ) 1 2 -D log(<unk> 2 0 ) + log(det( <unk>0 )) - 1 <unk> 2 0 u i T u i + (u i -<unk>i ) T <unk>-1 0 (u i -<unk>i ) = 1 2 DN log(<unk> 2 0 ) -N log(det( <unk>0 )) + N <unk> i=1 1 2 1 <unk> 2 0 E q(u i ) [u i T u i ] -E q(u i ) [(u i -<unk>i ) T <unk>-1 0 (u i -<unk>i )] = 1 2 DN log(<unk> 2 0 ) -N log(det( <unk>0 )) + N <unk> i=1 1 2<unk> 2 0 Var(u i ) + E q(u i ) [u i ] 2 - 1 2 N D = 1 2 DN log(<unk> 2 0 ) -N log(det( <unk>0 )) + N tr( <unk>0 ) 2<unk> 2 0 + N i=1 <unk>i T <unk>i 2<unk> 2 0 - 1 2 N D M <unk> a=1 KL[q(v a )|| f (v a )] = 1 2 D M log(<unk> = N <unk> i=1 M <unk> a=1 y ia <unk>1 - D <unk> d=1 E [u 2 id ] + E [v 2 ad ] -2E [u id ]E [v ad ] = N <unk> i=1 M <unk> a=1 y ia <unk>1 - D <unk> d=1 V ar [u id ] + E [u id ] 2 + V ar [v ad ] + E [v ad ] 2 -2E [u id ]E [v ad ] = N <unk> i=1 M <unk> a=1 y ia <unk>1 -tr( <unk>0 ) -tr( <unk>1 ) -( <unk>i -<unk>a ) T ( <unk>i -<unk>a ) The last 2 expectations of the log functions can be simplified using Jensen's inequality and -u j ) T (u i -u j ))] = E q(U,V |Y I,Y I A ) exp -Z T (2 <unk>0 ) 1/2 + u T (2 <unk>0 ) 1/2 Z + u = E q(U,V |Y I,Y I A ) exp -Z T (2 <unk>0 )Z -2Z T (2 <unk>0 ) 1/2 u -u T u) = 1 <unk> 2<unk> <unk> exp -Z T 2 <unk>0 + 1 2 I Z -2Z T (2 <unk>0 ) 1/2 u -u T u dZ Now define Q = u(2 <unk>0 + 1 2 I) -1 (2 <unk>0 ) 1/2. Then the above integral becomes 1 <unk> 2<unk> <unk> exp -(Z -Q) T (2 <unk>0 + 1 2 I)(Z -Q) -u T u + u T (2 <unk>0 + 1 2 I) -1 (2 <unk>0 )u dZ = exp -u T u + u T (2 <unk>0 + 1 2 I) -1 (2 <unk>0 )u det(I + 4 <unk>0 ) -1 2 = exp -u T (I -(2 <unk>0 + 1 2 I) -1 (2 <unk>0 ))u det(I + 4 <unk>0 ) -1 2 = exp -u T (4 <unk>0 + I) -1 u det(I + 4 <unk>0 ) -1 2. The last line follows since for any two invertible matrices A and B, if A + B is also invertible, then by Henderson and Searle (1981) ( A + B) -1 = A -1 -A -1 B(I + A -1 B) -1 A -1. Letting B = 4 <unk>0 and A = I gives: E q(U,V |Y I,Y I A ) [exp(-(u i -u j ) T (u i -u j ))] = exp -( <unk>i -<unk> j ) T (I+4 <unk>0 ) -1 ( <unk>i -<unk> j ) det(I+4 <unk>0 ) -1 2 Recall u i, f Z (z) = 1 <unk> 2<unk> exp(-1 2 z T z). Consequently, we have u i -v a = ( <unk>0 + <unk>1 ) 1/2 Z + u. Therefore, we can reparameterize E q(U,V |Y I,Y I A ) [exp(-(u i -v a ) T (u i -v a ))] = E q(U,V |Y I,Y I A ) exp -Z T ( <unk>0 + <unk>1 ) 1/2 + u T ( <unk>0 + <unk>1 ) 1/2 Z + u = E q(U,V |Y I,Y I A ) exp -Z T ( <unk>0 + <unk>1 )Z -2Z T ( <unk>0 + <unk>1 ) 1/2 u -u T u) = 1 <unk> 2<unk> <unk> exp -Z T <unk>0 + <unk>1 + 1 2 I Z -2Z T ( <unk>0 + <unk>1 ) 1/2 u -u T u dZ Now define Q = u( <unk>0 + <unk>1 + 1 2 I) -1 ( <unk>0 + <unk>1 ) 1/2. Then the above integral becomes 1 <unk> 2<unk> <unk> exp -(Z -Q) T ( <unk>0 + <unk>1 + 1 2 I)(Z -Q) -u T u + u T ( <unk>0 + <unk>1 + 1 2 I) -1 ( <unk>0 + <unk>1 )u dZ = exp -u T u + u T ( <unk>0 + <unk>1 + 1 2 I) -1 (

In many application problems in social, behavioral, and economic sciences, researchers often have data on a social network among a group of individuals along with high dimensional multivariate measurements for each individual. To analyze such networked data structures, we propose a joint Attribute and Person Latent Space Model (APLSM) that summarizes information from the social network and the multiple attribute measurements in a person-attribute joint latent space. We develop a Variational Bayesian Expectation-Maximization estimation algorithm to estimate the posterior distribution of the attribute and person locations in the joint latent space. This methodology allows for effective integration, informative visualization, and prediction of social networks and high dimensional attribute measurements. Using APLSM, we explore the inner workings of the French financial elites based on their social networks and their career, political views, and social status. We observe a division in the social circles of the French elites in accordance with the differences in their individual characteristics.

+ u = E q(U,V |Y I,Y I A ) exp -Z T ( <unk>0 + <unk>1 )Z -2Z T ( <unk>0 + <unk>1 ) 1/2 u -u T u) = 1 <unk> 2<unk> <unk> exp -Z T <unk>0 + <unk>1 + 1 2 I Z -2Z T ( <unk>0 + <unk>1 ) 1/2 u -u T u dZ Now define Q = u( <unk>0 + <unk>1 + 1 2 I) -1 ( <unk>0 + <unk>1 ) 1/2. Then the above integral becomes 1 <unk> 2<unk> <unk> exp -(Z -Q) T ( <unk>0 + <unk>1 + 1 2 I)(Z -Q) -u T u + u T ( <unk>0 + <unk>1 + 1 2 I) -1 ( <unk>0 + <unk>1 )u dZ = exp -u T u + u T ( <unk>0 + <unk>1 + 1 2 I) -1 ( <unk>0 + <unk>1 )u det(I + 2 <unk>0 + 2 <unk>1 ) -1 2 = exp -u T (I -( <unk>0 + <unk>1 + 1 2 I) -1 ( <unk>0 + <unk>1 ))u det(I + 2 <unk>0 + 2 <unk>1 ) -1 2 = exp -u T (I + 2 <unk>0 + 2 <unk>1 ) -1 u det(I + 2 <unk>0 + 2 <unk>1 ) -1 2. The last line follows since for any two invertible matrices A and B, if A + B is also invertible, then by Henderson and Searle (1981) ( A + B) -1 = A -1 -A -1 B(I + A -1 B) -1 A -1. Letting A = I and B = 2 <unk>0 + 2 <unk>1 gives: E q(U,V |Y I,Y I A ) [exp(-(u i -v a ) T (u i -v a ))] = exp -( <unk>i -<unk> a ) T (I+2 <unk>0 +2 <unk>1 ) -1 ( <unk>i -<unk> a ) det(I+2 <unk>0 +2 <unk>1 ) -1 2 Finally, the Kullback-Leiber divergence between the variational posterior and the true posterior is KL[q(U, V |Y I, Y I A )|| f (U, V |Y I, Y I A )] <unk> 1 2 DN log(<unk> 2 0 ) -N log(det( <unk>0 )) + N tr( <unk>0 ) 2<unk> 2 0 + N i=1 <unk>i T <unk>i 2<unk> 2 0 - 1 2 N D + 1 2 D M log(<unk> 2 1 ) -M log(det( <unk>1 )) + M tr( <unk>1 ) 2<unk> 2 1 + M a=1 <unk>a T <unk>a 2<unk> 2 1 - 1 2 M D - N <unk> i=1 M <unk> a=1 y ia <unk>1 -tr( <unk>0 ) -tr( <unk>1 ) -( <unk>i -<unk>a ) T ( <unk>i -<unk>a ) - N <unk> i=1 N <unk> j=1, j<unk>i y i j <unk>0 -2 tr( <unk>0 ) -( <unk>i -<unk> j ) T ( <unk>i -<unk> j ) + N <unk> i=1 M <unk> a=1 log 1 + exp( <unk>1 ) det(I + 2 <unk>0 + 2 <unk>1 ) 1 2 exp -( <unk>i -<unk>a ) T (I + 2 <unk>0 + 2 <unk>1 ) -1 ( <unk>i -<unk>a ) + N <unk> i=1 N <unk> j=1, j<unk>i log 1 + exp( <unk>0 ) det(I + 4 <unk>0 ) 1/2 exp -( <unk>i -<unk> j ) T (I + 4 <unk>0 ) -1 ( <unk>i -<unk> j ) + <unk> N i=1 <unk>i T <unk>i 2<unk> 2 0 - N <unk> i=1 M <unk> a=1 y ia <unk>1 -tr( <unk>0 ) -tr( <unk>1 ) -( <unk>i -<unk>a ) T ( <unk>i -<unk>a ) - N <unk> i=1 N <unk> j=1, j<unk>i y i j <unk>0 -2 tr( <unk>0 ) -( <unk>i -<unk> j ) T ( <unk>i -<unk> j ) + N <unk> i=1 M <unk> a=1 log 1 + exp( <unk>1 ) det(I + 2 <unk>0 + 2 <unk>1 ) 1 2 exp -( <unk>i -<unk>a ) T (I + 2 <unk>0 + 2 <unk>1 ) -1 ( <unk>i -<unk>a ) + N <unk>i=1 --- Supplementary Materials The supplementary materials contain detailed derivations of the variational bayesian EM algorithm for the proposed model, the APLSM, parts of which are used to estimate the BLSM.

In many application problems in social, behavioral, and economic sciences, researchers often have data on a social network among a group of individuals along with high dimensional multivariate measurements for each individual. To analyze such networked data structures, we propose a joint Attribute and Person Latent Space Model (APLSM) that summarizes information from the social network and the multiple attribute measurements in a person-attribute joint latent space. We develop a Variational Bayesian Expectation-Maximization estimation algorithm to estimate the posterior distribution of the attribute and person locations in the joint latent space. This methodology allows for effective integration, informative visualization, and prediction of social networks and high dimensional attribute measurements. Using APLSM, we explore the inner workings of the French financial elites based on their social networks and their career, political views, and social status. We observe a division in the social circles of the French elites in accordance with the differences in their individual characteristics.

In today's "permanently online, permanently connected" world (Vorderer et al., 2017), people increasingly make purposeful efforts to disconnect, such as by limiting their connectivity or time spent on digital media (Nguyen et al., 2022). For instance, individuals might take "digital detoxes" and use tech-based solutions such as apps and device features to unplug (Nguyen, 2021). Societal norms toward disconnection might also develop, as social groups and organizations implement guidelines and policies around digital media use (Vanden Abeele & Nguyen, 2023). Research reveals that well-being and health considerations are one of the central reasons why people decide to place limits on their digital media use or to "disconnect" (Jorge, 2019;Nguyen, 2023). Yet, in current scholarship on digital media use and its impact on subjective well-being, the role of everyday digital media practices that are geared toward limiting connectivity remains understudied. Exceptions to this are intervention studies where people are asked to take breaks from social media or mobile devices, but these do not reflect people's everyday lived experiences (e.g., Hall et al., 2021; for a systematic review, see Radtke et al., 2021). As such, research has not yet been able to address to what extent disconnection is able to restore people's sense of well-being, which is perceived to be threatened by digital media use (Nguyen, 2023). Note that ample research has been conducted on the relationship between digital media use and subjective well-being, finding mostly none to small effects (for a review, see Meier & Reinecke, 2020; for a discussion, see Orben, 2020), but individual perceptions that digital media affects well-being negatively still remain (e.g., Comparis, 2018;Kantar, 2018). Since well-being considerations are one of the key motivations to disconnect (Nguyen, 2023), it is important to examine if disconnection-that is, deliberately limiting the use of digital media-is an effective strategy to restore such perceptions of reduced well-being. Another important question is to what extent digital inequalities shape the relationship between digital disconnection and subjective well-being. In recent years, scholars have voiced concerns about how digital inequalities may extend to the realm of digital disconnection, putting forward the notion that some might be better able to manage the abundant digital information and communication environment than others (e.g., Gui & Bu <unk>chi, 2021;Hargittai & Micheli, 2019;Nguyen, 2021;Nguyen & Hargittai, 2023). In a world where connection is the default, disconnection may be reserved for the more digitally and socioeconomically privileged, who are able to "afford" going offline (Beattie & Cassidy, 2020;Bu <unk>chi et al., 2019). With many disconnective features being built into device and app interfaces (e.g., Apple Screen Time, "Focus" mode), it is likely that these will be more accessible to those who are digitally skilled enough to utilize them (Nguyen, 2021). Overall, digital inequality research has shown that more skilled digital media users are more likely to experience benefits from their digital media uses, for instance, in terms of increased social capital or well-being (Hofer et al., 2019;Nguyen et al., 2020). As such, an important question is whether and how digital skills play a moderating role in the relationship between everyday digital disconnection practices and subjective well-being. In this preregistered study, we examine the implications of people's everyday disconnection practices for their sense of well-being (i.e., affective, social, and cognitive well-being), and how digital skills play a role in this. We aim to explore people's broader digital media repertoire instead of only focusing on specific devices or platforms, as digital media use and henceforth disconnection likely does not happen in an isolated context (e.g., Nguyen et al., 2021). Regarding digital disconnection, we aim to capture the everyday practices in which people engage while maintaining connectivity-but without going completely offline for longer periods of time, such as in the case of digital detoxes where users take longer breaks (e.g., weeks and months) from mobile or social media (Baym et al., 2020;Franks et al., 2018;Radtke et al., 2021). To capture these everyday disconnection practices in their natural setting-as opposed to intervention studies that "force" participants to disconnect-this study employs an experience sampling method (ESM) design where digital media users are surveyed six times throughout the day over the course of a week to capture their possible disconnection practices as they occur naturally. The unique advantage of ESM in the context of this study is that it can capture (a) the nuanced disconnection practices in which people engage throughout the day (e.g., putting the phone with the screen down for several minutes; making a mental note not to visit certain apps or websites) and (b) the momentary, likely short-lived, effects of such nuanced disconnection practices on subjective well-being. --- Theorizing the relationship between disconnection and subjective well-being Digital media play an essential role in organizing everyday life, maintaining social relationships, and fulfilling leisure time, and as such may be important for nourishing affective, cognitive, and social well-being outcomes (Meier & Reinecke, 2020). It is not surprising that decades of research have been concerned with the question of how digital media use affects people's subjective well-being (for a review, see Meier & Reinecke, 2020). With respect to the impact of digital disconnection on various dimensions of subjective well-being, there are several hypotheses to consider. As a starting point, we first discuss existing theories about the relationship between digital media use and well-being, and then elaborate on what this would mean for potential effects of disconnection on different well-being outcomes. In this study, we define digital disconnection as nuanced acts of deliberately limiting one's digital media use (e.g., putting devices away, making mental notes not to look at devices, turning off notifications or the internet). Overall, it is important to note that digital disconnection and its potential consequences for well-being might differ depending on geographic and cultural context (Bozan & Trere <unk>, 2023;Trere <unk>, 2021). Much of the disconnection literature so far (for a review, see Nassen et al., 2023), as this study, has been focused on Western, educated, industrialized, rich, and democratic countries (Henrich et al., 2010), where dis/connectivity for many is likely a choice rather than a necessity (also see the Discussion and Conclusion section). On the one hand, scholarship has theorized digital media use to have a positive effect on well-being. For instance, digital media such as smartphones can help young people develop a sense of autonomy (Schnauber-Stockmann et al., 2021) and can support older adults in living an autonomous life (Abascal & Civit, 2000), which are fundamental human needs that contribute to overall well-being (Ryan & Deci, 2001). With respect to digital communication, researchers have suggested that both the quantity and quality of human communication can be improved through use of technology, which in turn contributes to greater social connection and overall well-being (Dienlin et al., 2017;Valkenburg & Peter, 2007). If we think of digital media use, that is, connection, as having positive effects on subjective well-being, then one hypothesis could be that disconnection interferes with this relationship and thus negatively impacts well-being. On the other hand, digital media use may also have negative consequences for subjective well-being. For instance, a well-known hypothesis is that of "displacement" (Dienlin et al., 2017;Kraut et al., 1998;Valkenburg & Peter, 2007), whereby time spent on digital media detracts from time that could be spent on other potentially more meaningful activities, including in-person social ones. For instance, in the case of problematic social media use-when people feel a loss of control over their social media use or experience "overconnection"-research has shown that it can lead to increased loneliness (Marttila et al., 2021) and is thus detrimental to one's well-being. Digital media have also been experienced as distracting, leading to unwanted effects such as procrastination (Hinsch & Sheldon, 2013;Meier, 2021). Disconnection, then, could be a solution to such perceived negative effects and allow people to spend their free time on other activities that are perceived to be more meaningful, all the while regaining a sense of control over their digital media use (Aranda & Baig, 2018;Franks et al., 2018;Nguyen, 2021). So far, in theorizing the relationship between digital disconnection and subjective well-being, we have given examples in which we think of disconnection as the reverse of connection, and likewise the effects it may have on subjective well-being. While this might seem logical and straightforward, we argue that effects of digital disconnection on well-being are not necessarily the opposite of those of digital media use. Unlike digital media use, which is often described as automatic, habitual (Bayer et al., 2022;Giannakos et al., 2013;Wohn, 2012) and perhaps even as mindless (Baym et al., 2020;Schellewald, 2021), when it comes to disconnection practices, people engage in them more consciously and in a goal-oriented manner (Aranda & Baig, 2018;Franks et al., 2018;Nguyen, 2021). Through disconnecting, people often hope to achieve something specific, such as being more present in offline activities, not being distracted when wanting to concentrate on an activity, or escaping the social pressure to be online (Nguyen, 2023). These observations are in line with Communication Bond Belong theory, which posits that the motivation to engage in social interactions decreases when the need to belong is fulfilled, and vice versa, because social interactions come at the expense of social energy, which people need to preserve (Hall & Davis, 2017). Recently, scholars have linked this theory to the concept of digital solitude (Campbell & Ross, 2022), suggesting that when people actively decide to disconnect, this can help them regain their social energy (Ross et al., 2023). From the various points of view discussed, it would make sense to expect that, overall, people's disconnection practices have a positive impact on well-being in general. Nonetheless, it is important to note that digital media use itself (rather than disconnection) can also be practiced as a moment of digital solitude, for instance when someone uses social media by oneself for a moment of relaxation or to cope with offline stress (Keessen, 2023;Wolfers & Utz, 2022), and so the relationship of disconnection to well-being may not be that straightforward. What does existing empirical scholarship on digital disconnection have to say about its effects on subjective well-being? Research on the relationship between disconnection and subjective well-being has so far mostly been of experimental nature, where participants are instructed to abstain from social media or their smartphones for a certain time period. Overall, the results of such studies are mixed: Some find positive effects on affective, cognitive, and social well-being outcomes, while others find no such effects, or even negative ones (for a systematic review, see Radtke et al., 2021). In one of the most comprehensive studies to date, Hall et al. (2021) randomly assigned participants to one of five experimental conditions in which they were either asked to make no changes to their social media use or take a break from social media ranging from 1 to 4 weeks. Relying on daily diary assessments, the authors found no effect of abstention nor the duration of abstinence on affective, social, or cognitive dimensions of subjective wellbeing. Aside from intervention studies, there is also research that relies on participants' recall of their disconnection experiences. For instance, in a study where researchers collected open survey responses, Facebook users reported that after a time of deactivation, they returned to be more mindful users of the platform (Baym et al., 2020). Cross-sectional survey research has shown that intended unavailability to communicate with others (either in person or digitally) can increase overall wellbeing, but only for users who are more digitally connected than others (Ross et al., 2023). An interview study found that taking breaks from social media could benefit people's sense of health and well-being, although this was not the case for all interviewed participants (Nguyen, 2021). On the whole, while experimental studies show mixed effects of forced disconnection from social media/mobile phones on subjective wellbeing (e.g., Radtke et al., 2021), interview studies that ask about voluntary disconnection experiences in retrospect report overall benefits for perceived well-being (Aranda & Baig, 2018;Baym et al., 2020;Franks et al., 2018;Nguyen et al., 2020). This leaves the question of whether method of analysis may influence how disconnection and well-being are linked. Observational studies that focus on the effects of disconnection strategies implemented in day-to-day life on subjective well-being (unlike forced "detoxes" where people opt out of social media for a certain time period due to researcher prompting) are less present in the literature. Yet, such everyday disconnection practices are quite common and, as such, are ripe for investigation. For instance, research from Switzerland indicates that two-thirds of digital media users engage in at least one strategy to disconnect (e.g., turning off notifications, using do-not-disturb-functions, or putting digital devices away), with health and well-being among the most prominent reasons listed for doing so (Nguyen et al., 2022). In an Austrian study of people ages 18-35, almost half used screen time apps on their smartphone (Schmuck, 2020). Moreover, among people who do not use screen time apps, social network site use was positively associated with problematic smartphone use, and consequently, lower subjective well-being, while there was no such link among screen time app users (Schmuck, 2020). As such, it could be that engaging in efforts to regulate one's media use can mitigate potential negative effects of digital media use on subjective well-being. While these studies are helpful as a first step, more research is needed to deepen our understanding of the possible effectiveness of everyday disconnection practices for subjective well-being. --- Momentary effects of disconnection on subjective well-being Given that people often engage in disconnection practices consciously and with a specific goal in mind (e.g., being more present in offline activities; Aranda & Baig, 2018;Franks et al., 2018;Nguyen, 2021), we expect that the effects on affective and cognitive well-being overall will be positive. However, when it comes to social well-being, existing empirical work seems to suggest that in the short-term, disconnection can lead to diminished feelings of connectedness. Nguyen (2021) found this to be the case in an interview study with 30 adults who had taken a break from social media at least once. People's overall, long-term reflections were positive and reflected increased well-being, whereas short-term, they had reported negative effects on social well-being, such as fear of missing out and restlessness when they decided to deactivate Facebook or put their mobile phone away (Nguyen, 2021). When comparing experimental studies in which people were asked to disconnect from social media, a one-week intervention yielded an increase in social connectedness (Brown & Kuss, 2020), while another study with a shorter two-day intervention led to lower connectedness (Sheldon et al., 2011), although the latter finding likely does not reflect momentary effects. A comprehensive study which compared different lengths of social media breaks (ranging from 0 to 4 weeks) found no effects of the duration of abstinence on loneliness (Hall et al., 2021). Research using the ESM to examine communication between close ties (e.g., couples) has shown that connected availability-the perception that a partner is continuously available digitally-benefits subjective well-being (Taylor & Bazarova, 2021). Theoretically, disconnection then could be detrimental to feelings of social connectedness-at least in the short term, as it interferes with this state of "connected presence" (Christensen, 2009;Licoppe, 2004) to others at all times. In the context of in-person social interactions, however, phubbing-interrupting a social situation through smartphone use-can lead to detrimental well-being (Nu<unk>ez & Radtke, 2023). Disconnection, then, could increase especially perceived social well-being. As an observational study showed, people experienced more intimate conversations when smartphones were used less (Vanden Abeele et al., 2019). This study aims to examine the momentary effects of people's disconnection practices throughout the day on three dimensions of subjective well-being: affective (i.e., positive/ negative mood), cognitive (i.e., life satisfaction), and social (i.e., social connectedness), as these can be considered the core elements of general well-being (Andrews & McKennell, 1980;Diener et al., 2009). In the case of such everyday disconnection practices (e.g., putting the phone away or turning off the internet for a short time period), effects on subjective well-being indicators may be short-lived and limited to how one feels in the moment, thus requiring momentary assessment methods to capture. Our aim is to include people's broader digital media repertoire, thus focusing on devices that grant access to the internet, such as a smartphone, laptop/ computer, or tablet, and the apps and programs that one can use through them. Inspired by current theorizing and empirical work in the area of digital media use, digital disconnection, and well-being, we expect that engaging in disconnection practices will benefit affective and cognitive well-being in particular. However, for social well-being, we expect negative effects in the short-term. As such, we formulate the following hypotheses: 1 H1: Everyday disconnection practices are related to higher perceived affective well-being. H2: Everyday disconnection practices are related to higher perceived cognitive well-being. H3: Everyday disconnection practices are related to lower perceived social well-being. --- Digital inequalities in disconnection: the role of digital skills Scholars have suggested that socio-digital inequalities exist around people's disconnection practices and have highlighted the importance of digital skills-that is, the ability to use technology effectively and efficiently (Hargittai & Hsieh, 2012)in managing the abundant information and communication that people are exposed to through digital media (Gui et al., 2017;Hargittai et al., 2012;Hargittai & Micheli, 2019). A survey of American internet users revealed that people with greater digital skills were more likely to use strategies to limit their digital media use during the COVID-19 pandemic (Nguyen & Hargittai, 2023). Qualitative interview studies have also reported that digital skills are important for being able to use features of technology to limit connectivity (Nguyen, 2021;Nguyen et al., 2021). Examples of such strategies include the use of screen time apps to track and set limits to one's digital media use, as well as the use of app and device settings to manage one's online availability or exposure to online information and communication. Overall, since evidence of an association between digital skills and disconnection behaviors is still scarce, a first question that we will explore in this article is whether people's digital skills predict the extent to which they engage in nuanced disconnection practices throughout the day. While previous research has established that digital inequalities in people's disconnection behaviors exist, the extended question of whether this results in differentiated benefits derived from such deliberate non-use remains unanswered to date. This focus on disparities in the benefits that people are able to derive from technology use is an important aspect of digital inequality (Hargittai, 2008): Digital inequality research suggests that in order for people to reap the benefits from using digital media, people need to possess the skills needed to capitalize on such use (DiMaggio et al., 2004). For instance, research on digital media use and social and cognitive wellbeing in older adults reveals a positive relationship, which was stronger for those with greater digital skills (Hofer et al., 2019;Nguyen et al., 2020), indicating that digital skills are an important moderator of media effects. This suggests that people with adequate digital skills can mobilize technology use for greater benefits-meaning that more digitally privileged people might disproportionally profit from engaging in digital activities. Given the importance of digital skills for engaging in disconnection practices (Nguyen & Hargittai, 2023), the question of whether such differences also extend to the benefits that people potentially derive from such disconnection is a highly relevant one. In this study, we also explore the role of digital skills for the benefits that people reap from digital disconnection in terms of their subjective well-being. Here, we ask the following research questions: RQ1: Do digital skills predict people's engagement in disconnection practices? 2 RQ2: Do digital skills moderate the relationships between digital disconnection and (a) affective, (b) cognitive, and (c) social well-being? --- Methods In this preregistered study, we use ESM (for more information about this method, see Myin-Germeys & Kuppens, 2022) to examine the effects of everyday disconnection practices on three different dimensions of subjective well-being: affective, cognitive, and social. We also examine how these relationships are dependent on people's digital skills. ESM is an appropriate method to examine the occurrence of everyday disconnection practices that happen throughout the day, as well as the immediate, momentary, short-lived effects that it may have on subjective well-being. The hypotheses, research questions, design, and analytical plan were preregistered (https://osf.io/6kw9c). --- Design and sampling plan We recruited Dutch participants 18 years and older through a market research company in March and April 2022. We aimed for a somewhat equal distribution of gender, age, and education levels, which the research company took into account in sending out invitations. Prospective participants received an invitation e-mail from the research company with the study information letter and a link to the informed consent form and intake questionnaire. In the study information letter, we explained the study procedure and that people needed to possess an Android smartphone and install the mobile experience sampling app (movisensXS) to participate. On day 1, participants received an intake questionnaire to measure sociodemographics and general digital media use. Through this intake survey, participants also received step-bystep instructions to install the movisensXS app. We also gaveinstructions to change device settings to receive questionnaire notifications from the app, even in "do-not-disturb" mode (this is a default option of the app, which is needed so that it can function properly). On days 2-8 (i.e., 7 consecutive days), participants received six questionnaires each day through movisensXS at semirandom time intervals between 8:30 am and 9:00 pm (see exact sampling schedule in preregistration). Given that our questions (see "Daily ESM Questionnaire") asked about the preceding 2-hr time window for several questions, we ensured that all questionnaire prompts were at least 2 hr apart to avoid overlap in the 2-hr time windows. In total, participants received 42 questionnaires through the app. For most participants (99 out of 105) we needed to extend the ESM part of the study (by 1 day n 1<unk>4 64; 2 days n 1<unk>4 12; 3 days n 1<unk>4 10; 4 days n 1<unk>4 6; 5 days n 1<unk>4 5; 6 days n 1<unk>4 2), either because they had missed questionnaires on the first days due to unforeseen technical issues or because of personal reasons such as illness (see under the Analyses section that the day of measurement did not influence the results). For those willing to continue their participation in the study in spite of this, we sampled until they filled out at least 33 questionnaires (which was one of our inclusion criteria in the preregistration based on power calculations). On the 9th day, participants received a final outtake questionnaire (these data are not part of this article). Participants received a financial compensation of 25 Euros upon completion of the study if they filled out the intake and final questionnaire, and if they filled out at least 80% of the ESM questionnaires (i.e., a minimum of 33), with at least one each day. Participants were instructed to fill out the questionnaire as soon as possible after receiving a notification, but it stayed available for 30 min after the first prompt. Alternatively, participants could delay the notification up until 25 min later, or they could skip the questionnaire. Participants received four reminder notifications if they had not filled out the questionnaire in those 30 min. Given this study's focus on disconnection practices, it is important to note that the mobile experience app also worked offline (i.e., when there was no internet connection). The app would then upload the results to the server once the participant again had an active internet connection. The irony of using mobile media to study disconnection practices is not lost on the authors. We nonetheless believe that this is still an appropriate method to study this behavior, as we do not expect to include participants who disconnect from their phone for longer periods of time (i.e., days, weeks). Instead, we focus on the brief, nuanced disconnection practices that active digital media users engage in throughout the day. Extended disconnection is quite rare, and likely limited to a select group of people who are extremely serious about disconnecting (Nguyen, 2021). --- Participants In total, 151 people signed through the experience sampling app, of which n 1<unk>4 14 never filled out any questionnaires, and n 1<unk>4 32 ended their participation early. Our final sample consisted of 105 participants, of which two-thirds were female (68.6%). The average age of participants was 40.10 years (SD 1<unk>4 13.10) and ranged from 19 to 75 years. The sample included both lower educated (50.5%) and higher educated participants who had earned at least a Bachelor's degree (49.5%). In total, we sent 5055 ESM questionnaires to our 105 participants, of which they filled out 4028 (79.7%). On average, participants completed 38.4 ESM questionnaires (range 32-42). Note that we had an inclusion criterion of filling out at least 33 ESM questionnaires to be included in the final sample of our study (for more detailed information on the rationale behind this decision, see our preregistration). However, as two participants started, but did not finish the 33rd questionnaire, we ended up including all participants who completed 32 questionnaires. --- Baseline questionnaire --- Sociodemographics We measured age by asking for people's birth year and subtracting that from 2022. We asked about gender (female, male, other), which we recoded into a female category (vs. other). We measured education level through nine categories, ranging from no formal education to having obtained an academic degree, which we recoded into a dichotomous variable reflecting higher education level (i.e., university versus all lower education categories). --- General digital media use We asked about people's use of the internet and their smartphones separately, distinguishing between weekdays and the weekend, which resulted in four items. The question was: "On an average [weekday/Saturday or Sunday], how often do you use [the internet, either on a computer, tablet, or smartphone/your smartphone]?". We calculated the average use per day by taking the sum of the weekday answer * 5 days and the weekend answer * 2 days, and dividing this by 7 days, resulting in two variables separately reflecting average internet use and smartphone use per day. On average, participants used the internet for 4.20 hr per day (SD 1<unk>4 1.43), and their smartphone for 3.35 hr per day (SD 1<unk>4 1.61). --- Digital skills For digital skills, we used an established and validated index to measure people's know-how about the internet and social media (Hargittai, 2020;Hargittai & Hsieh, 2012). This selfreported index has partly been validated against people's actual performance of skills (Hargittai, 2005). Respondents reported their understanding of 12 internet and social mediarelated items on a 1-5 point scale ranging from no understanding to full understanding (e.g., "PDF," "cache," "followers," "tagging"). We took the average of the items as the digital skills score (Cronbach's a 1<unk>4 0.87). The average level of digital skills was relatively high, but people with lower levels of skills were also represented in the sample (M 1<unk>4 3.88, SD 1<unk>4 0.78, range 1-5). --- Daily ESM questionnaire --- Disconnection We pretested the question and answer options for measuring disconnection (n 1<unk>4 9) for clarity, comprehensibility, and completeness during a pilot of the study procedure and made several adjustments before finalizing the version presented here. To measure disconnection practices in everyday life, we asked people: "In the past two hours, have you done any of the following deliberately to take a break from digital media? Check all that apply." We communicated to participants that digital media entailed "devices that grant access to the internet, such as a smartphone, laptop/computer, or tablet." The nine answer options were: (a) put my mobile phone or other digital media away; (b) put my mobile phone or other digital media with the screen facing down; (c) told myself not to look at my mobile phone or other digital media; (d) told myself not to look at certain websites or apps (for instance, email, news, and social media); (e) closed programs, apps, and websites (for instance, email, news, and social media); f) turned off notifications (for instance from email, news, and social media); (g) turned off the internet (for instance, flight mode, "donot-disturb" function); (h) took a break from digital media in another way; and (i) limited digital media use without any special approach. At the end of the list, we included a 10th option for people to report that they had not disconnected in the past 2 hr: "I have NOT taken a break from digital media in the past two hours." Confirmatory factor analyses (using the R package lavaan; Rosseel et al., 2023) relying on tetrachoric correlations suitable for dichotomous variables showed an acceptable model fit (v 2 (27, N 1<unk>4 3972) 1<unk>4 308.92, P <unk>.001, CFI 1<unk>4 0.800, RMSEA 1<unk>4 0.051; Hooper et al., 2008), justifying our decision to combine the nine items for measuring disconnection into one index. Thus, for our analyses, we created a dichotomous variable reflecting whether people had disconnected in the past 2 hr (versus not). If people indicated that they had used one of the disconnection strategies while simultaneously reporting that they had not disconnected in the past two hours, we recoded their responses as missing since those answers are logically misaligned. If people had indicated that they had disconnected, we also collected data on whom they were with when they disconnected. Answer options were: alone, partner, child(ren), other family, friend(s), colleague(s), and other; people could select multiple answer options. For the analyses, we created a dichotomous variable for each type of social tie. We also created another dichotomous variable to reflect whether people had disconnected in the copresence of other(s) (1) versus while alone (0). --- Subjective well-being We focused on three dimensions of subjective well-being, namely affective, cognitive, and social well-being. We measured affective and social well-being at every occasion, that is, six times per day. Given that we expected less variability during the day in cognitive well-being than other dimensions of subjective well-being (de Vries et al., 2021), we asked this only once: in the last questionnaire of each day. We used oneitem measures for these constructs, which is common in experience sampling research to lower participant burden. We conceptualized affective well-being as people's experience of negative and positive emotions or mood. To capture affective well-being, we asked: "How do you feel at this moment?", with answer options ranging from 1 "Very negative" to 7 "Very positive" (de Vries et al., 2021). We conceptualized cognitive well-being as one's general satisfaction with life, and asked: "When you look at the past day, how satisfied are you with your life at this moment?" Participants could answer on a slider ranging from 1 "Very unsatisfied" to 100 "Very satisfied" (de Vries et al., 2021). We used a slider with a larger range, as this has been shown to be a reliable and valid method to assess constructs related to overall quality of life, such as life satisfaction (Benrud-Larson et al., 2005;de Boer et al., 2004). Finally, we conceptualized social wellbeing as people's sense of social connectedness, where we asked: "How socially connected do you feel at this moment?," with answer options ranging from 1 "Not at all" to 7 "Very much" (Arjmand et al., 2021). --- Analyses We first conducted exploratory analyses to look into the prevalence of disconnection practices across our sample, as well as how sociodemographics, digital experiences, and digital skills (RQ1) related to disconnection practices. Here, we used mixed effects logistic regression to model binary outcome variables, with a random intercept varying by participant. The outcome variables were disconnection behavior in the past two hours (yes/no), and whom they were with when they disconnected (dichotomous variable for each type of social company: alone, partner, child(ren), other family, friend(s), colleague(s), and other). Next, we tested the hypotheses with multilevel linear regression modeling to account for the nested structure of the data. We estimated two-level models in which repeated measurements of digital disconnection (level 1) are nested within individuals (level 2). To test H1, H2, and H3, we estimated separate models for each dependent variable (affective, cognitive, and social well-being). For affective and social well-being as outcomes, which were measured at six times daily, we estimated a fixed-effects model with disconnection as predictor. For cognitive well-being, which was measured once per day, we estimated a fixed-effects model with the sum score of total disconnection during the day as predictor. We conducted additional exploratory analyses to examine if the within-person effects of

Drawing on theories of digital media (non-)use and well-being, this study examines how voluntary disconnection relates to subjective well-being and what role digital skills play in this relationship. We rely on mobile experience sampling methods to link nuanced disconnection practices throughout the day (e.g., putting screen devices away and muting notifications) with momentary experiences of well-being. We collected 4,028 responses from 105 mobile media users over the course of one week. Multilevel regression analyses revealed neither significant within-person effects of disconnection on affective well-being, social connectedness, or life satisfaction, nor a significant moderation effect of digital skills. Exploratory analyses, however, show that effects of disconnection on well-being vary greatly across participants, and that effects are dependent on whether one disconnects in the physical copresence of others. Our study offers a refined perspective on the consequences, or lack thereof, of deliberate non-use of technology in the digital age.

we used mixed effects logistic regression to model binary outcome variables, with a random intercept varying by participant. The outcome variables were disconnection behavior in the past two hours (yes/no), and whom they were with when they disconnected (dichotomous variable for each type of social company: alone, partner, child(ren), other family, friend(s), colleague(s), and other). Next, we tested the hypotheses with multilevel linear regression modeling to account for the nested structure of the data. We estimated two-level models in which repeated measurements of digital disconnection (level 1) are nested within individuals (level 2). To test H1, H2, and H3, we estimated separate models for each dependent variable (affective, cognitive, and social well-being). For affective and social well-being as outcomes, which were measured at six times daily, we estimated a fixed-effects model with disconnection as predictor. For cognitive well-being, which was measured once per day, we estimated a fixed-effects model with the sum score of total disconnection during the day as predictor. We conducted additional exploratory analyses to examine if the within-person effects of digital disconnection on subjective well-being (H1-H3) differ across individuals. To do so, we extended each model with a random slope and tested if the model fit improved when doing so. To answer RQ2, we extended each of the aforementioned models with an interaction term between digital skills and disconnection. We also conducted all analyses with age, gender, education level, and general digital media use as control variables. Since these were not significant, we report the analyses without them. We also considered the measurement day type (weekday versus weekend) and whether it was the 1st-6th questionnaire of the day, but given that these showed no effect on the dependent variables, we excluded them from further analyses. The time-level variables (level 1) were personmean centered as we are interested in the within-subject effects. The individual-level variables of digital skills and general media use (level 2) were grand-mean centered. All analyses were conducted in R, using package lme4 (Bates et al., 2022) and lmerTest (Kuznetsova et al., 2020). --- Results --- Disconnection practices (exploratory analyses and RQ1) Table 1 displays the prevalence of disconnection practices. In 39.5% (n 1<unk>4 1,567) of completed ESM questionnaires, participants indicated that they had used at least one of the listed disconnection strategies in the preceding two hours. Recall that the question asked whether participants had engaged in these actions "deliberately to take a break from digital media." The most commonly used strategies were "Put my mobile phone or other digital media away" (23.4%, n 1<unk>4 928), "Put my mobile phone or other digital media with the screen facing down" (7.4%, n 1<unk>4 295), and "Limited digital media use without any special approach" (7.6%, n 1<unk>4 303). Participants, when they had disconnected, used on average 1.36 strategies (SD 1<unk>4 0.77) with a range of 1-8 strategies. Specifically, in 75.4% (n 1<unk>4 1,182) of filled-out ESM questionnaires where people said they had disconnected, only one strategy was used, followed by 17.5% (n 1<unk>4 274) with two strategies, and 4.9% (n 1<unk>4 76) with three strategies. Next, we looked into the role of sociodemographics and digital skills in disconnection behavior, and the social context of disconnecting (Table 2). Mixed-effects logistic regression analysis revealed that gender, age, and education level did not relate to disconnection behavior over the course of one week. Digital skills (RQ1) were also not related to people's engagement in disconnection practices throughout the day. When looking into the social context in which people disconnected, we found significant age and gender patterns. Specifically, older age was related to greater odds of disconnecting around one's partner. We found no significant relationships between sociodemographics and disconnecting around children. On the other hand, older age was related to lower odds of disconnecting around other family members (besides one's partner and children), friends, and colleagues-suggesting that younger people are more likely to do so in these social contexts. Female participants also had lower odds of disconnecting while being with friends, as compared to male participants. --- Disconnection practices and subjective well-being (H1-H3) The results of the multilevel linear regression models examining the relationship between disconnection and well-being are detailed in Tables 345. We first estimated intercept-only models, where we find intra-class correlations (ICC) of r 1<unk>4 0.39 for affective well-being, r 1<unk>4 0.48 for life satisfaction, and r 1<unk>4 0.41 for social connectedness. In general, the larger the ICC, the lower the variability within a person (and thus momentary assessments become irrelevant), and the higher the variability between the persons. In this case, our ICCs mean that over half of the variability in the outcome variables is due to momentary fluctuations within persons. As such, by using multilevel models, we can better model the variation in the outcome variables by allowing for individual differences in comparison to a model that does not take into account the multilevel structure of the data. Next, we included disconnection behavior as a predictor in the models to test the hypotheses (Tables 345, Model 1). Here, we found no significant effects of disconnection behavior on affective well-being, life satisfaction, or social connectedness. Thus, when a participant disconnected more or less often than they would do on average, this did not change their level of subjective wellbeing. As such, our hypotheses are rejected. Additionally, we ran exploratory models with disconnection as a dichotomous variable (i.e., in its uncentered form). This allowed us to test whether a situation in which someone disconnected (yes/no)-rather than whether someone disconnected more on less than they normally would have-led to a change in their well-being. Here, we also find no significant effects of disconnection behavior on well-being outcomes. Finally, we also explored the possibility that there could be lagged effects of disconnection at one time point, and subjective well-being at the next time point (only for affective wellbeing and social connectedness, as life satisfaction was only assessed at the end of the day). These results also did not yield additional insights. --- Individual differences in the disconnectionsubjective-well-being relationship (exploratory analyses) Inspired by previous work in the domain of social media use and well-being (Valkenburg et al., 2021), we conducted exploratory analyses in which we accounted for random slopes to examine whether there are between-person differences in the within-person relationship on disconnection practices and subjective well-being. Extending the models for H1-H3 with random slopes (Tables 345, Model 3) significantly improved the model fit for predictions of affective well-being (X 2 (2) 1<unk>4 17.97, P <unk>.001) and social connectedness (X 2 (2) 1<unk>4 44.91, P <unk>.001), but not for life satisfaction (X 2 (2) 1<unk>4 2.84, P 1<unk>4.242). Figure 1 displays the random effects for the relationship between disconnection and (a) affective well-being and (b) social connectedness. For the relationship between disconnection and affective well-being (Figure 1a), b ranged from -.18 (negative) through.18 (positive), with 11 participants (10.5%) experiencing effects of b 1<unk>4 -.05 and lower, 73 people (69.5%) between b 1<unk>4 -.05 and.05, and 21 participants (20.0%) of b 1<unk>4.05 and higher. For the relationship between disconnection and social connectedness (Figure 1b) b ranged from -0.40 (negative) through 0.29 (positive), with 32 participants (30.5%) experiencing effects of b 1<unk>4 -.05 and lower, 57 people (54.3%) between b 1<unk>4 -0.05 and.05, and 16 participants (15.2%) of b 1<unk>4.05 and higher. Although the effects are small, our analyses reveal that the within-person relationship between disconnection and affective well-being, as well as disconnection and social connectedness, differs across people. --- The moderating role of digital skills (RQ2) To examine the possible moderating effect of digital skills on the relationship between disconnection and subjective wellbeing, we added a cross-level interaction term to the fixedeffects regression models (Tables 345, Model 2). We found no significant moderation effects, meaning that the withinperson relationship between disconnection and subjective well-being did not vary depending on people's level of digital Note: B 1<unk>4 unstandardized coefficient, b 1<unk>4 standardized coefficient, CI 1<unk>4 confidence interval; r 2 1<unk>4 variance, ICC 1<unk>4 intraclass correlation coefficient. * p <unk>.05, *** p <unk>.001. skills. As exploratory analyses (Tables 345, Model 4), we also ran models with cross-level interaction effects between disconnection and digital skills to examine if the varying slopes in the random-effects model (i.e., differences in effects of disconnection on well-being across people) could be explained by their different levels of digital skills. These analyses did not yield additional insights. --- The role of social context of disconnection (exploratory analyses) Finally, we also explored the role of the social context in which people are disconnecting (Tables 345, Model 5). One reason for disconnection might be the copresence of others and therefore an immediate desire to disconnect from technology. For these analyses, we focused on the subset of observations in which disconnection took place and used a dichotomous variable reflecting the physical copresence of others during a disconnection attempt (1 versus 0 being alone; see "Daily ESM Questionnaires" in the Methods section). Similar to H1-H3, we ran multilevel linear regression models examining the relationship between disconnection in the physical copresence of others (predictor) and subjective wellbeing (outcome). The analyses revealed significant main effects indicating that when people disconnected while around others, their affective well-being, social connectedness, and life satisfaction momentarily improved. --- Discussion and conclusion This preregistered study aimed to examine the effects of everyday disconnection practices on people's affective, cognitive, and social well-being, as well as the role of digital skills in this relationship. By employing mobile ESM, we were able to capture people's nuanced disconnection practices throughout the day (e.g., putting the phone away or with the screen down; deciding not to use certain apps or websites), as well as give insight into the momentary, short-lived effects of such deliberate non-use on subjective well-being. By applying this methodological approach to addressing this relationship, our work offers a refined and nuanced perspective on the implications, or lack thereof, of deliberate non-use of technology in a digital world. With our work, we contribute to the currently small, but rapidly growing, body of literature on digital disconnection and well-being. While well-being considerations are often central to people's decision to disconnect (Franks et al., 2018;Nguyen et al., 2022), our study found no within-person relationship between people's disconnection behaviors and subjective wellbeing outcomes overall. Our findings align with intervention (i.e., experimental) research that finds mixed effects of digital disconnection (in particular mobile phone and social media) on subjective well-being outcomes (Radtke et al., 2021). It also aligns with broader research on media use and subjective well-being, showing that effects of digital media use remain ambiguous. In the most comprehensive review on digital media use and well-being to date, Meier and Reinecke (2020) have disentangled the relationships between digital media use and well-being by distinguishing between different levels of analysis of digital communication channels and types of wellbeing indicators. Overall, scholarship on digital disconnection, digital media use, and subjective well-being so far suggests that the non-use and use of digital media can have both positive and negative associations with well-being, that the associations-if any-are often small in size, and that associations are dependent on the specific outcomes investigated, as well as on moderators and mediators (Meier & Reinecke, 2020). Our work shows that momentary deliberate disconnective acts-at least in the short-term-have no clear positive or negative effects on subjective well-being outcomes. One possible explanation for the lack of effects is that the nuanced disconnection practices examined in this study (e.g., putting digital devices with their screen down, telling oneself not to look on certain websites and apps) are not impactful enough to bring about change in one's subjective well-being throughout the day. Furthermore, in our study we treated any disconnection in the past two hours as a dichotomous variable, and unfortunately, we have no insight into the length of people's disconnection attempts (i.e., was it a matter of minutes or a full hour?). It could be, nonetheless, that increased engagement in these practices over time has beneficial effects on one's sense of well-being. This would align with qualitative research where in-depth interviews reveal that people's retrospective, global experiences of digital connection are positive and considered to be beneficial to people's wellbeing (Franks et al., 2018;Nguyen, 2021). Alternatively, it may be that a period of a week was not long enough to detect any changes in subjective well-being. While our power calculations-as outlined in our preregistration-demonstrate that we should have enough measurements over the course of seven days to detect small effects, it could be worthwhile to examine these within-person relationships over extended periods of time, for instance using measurement burst designs (Hofer, 2020). Our exploratory analyses showed that people vary vastly in the effects that disconnection has on their subjective wellbeing, ranging from negative to positive. Therefore, a second explanation for the lack of main effects is that this individual variation evens out the main effects. An important follow-up question for future research is what individual factors explain who experiences benefits from disconnecting and who experiences drawbacks from doing so. It could be that the impact of disconnection practices on subjective well-being outcomes is linked to one's motivations to disconnect. In-depth interviews with people who had taken an extended break from social media revealed that there were different motivations for doing so, and that if people had not been bothered by the impact of digital media use on their well-being to begin with, they also did not feel that their well-being improved after disconnecting (Nguyen, 2021). Related to people's motivations to disconnect, this study did not consider in what activity people engaged while they were disconnected. As reasons for disconnection can be diverse, one can imagine that some activities undertaken while disconnected could be draining (e.g., working, studying) and thus detrimental for well-being, while others could be relaxing and beneficial (e.g., reading a book, exercising, and spending time with friends). There are also numerous other reasons for disconnecting such as disinterest, privacy concerns, political activism, and health considerations (e.g., Baumer et al., 2013;Hesselberth, 2018;Nguyen et al., 2021;Nguyen, 2023). Furthermore, it is important to consider that momentary digital disconnection could be more influential on immediate outcomes such as distraction and procrastination than on overall well-being per se (Valkenburg, 2022). Future theoretical models of disconnection effects and empirical research could account for how motivations and behaviors shape the consequences that disconnection has for various outcomes, including both momentary and longer-term subjective well-being. Our study also revealed no moderating effect of digital skills in the relationship between digital disconnection and subjective well-being. We also did not find a relationship between digital skills and disconnection behavior, which is in contrast with earlier research suggesting the existence of socio-digital inequalities in disconnection practices (Gui & Bu <unk>chi, 2021;Nguyen, 2021;Nguyen & Hargittai, 2023). It could be that digital skills are particularly important for disconnection strategies that are highly technical, such as adjusting device and app settings (Nguyen, 2021). In our study, many of the strategies listed-and especially the most used strategies (e.g., putting screen devices away/with the screen down)-do not require advanced digital skills but perhaps rather high levels of self-regulation. For instance, research shows that people with lower self-control are more likely to respond to incoming notifications immediately (Berger et al., 2018), which suggests they find it more difficult to ignore them. As such, a valid question is to what extent one's self-control about remaining disconnected and not checking in on digital media is important for the success of one's disconnection attempt and the benefits that people reap from it in terms of their well-being (Vanden Abeele et al., 2022). Digital skills, then, likely matter more when it concerns more technologically advanced ways of disconnecting. Alternatively, it could be that our measure of digital skills (which asked about internet-and social media-related topics) does not reflect the skills needed to adjust device/app settings, although this would only be relevant for a few of the disconnection strategies listed (i.e., turn off notifications, turn off the internet), which were generally used less often. Moving forward, it could be fruitful to differentiate disconnection by device, platform, and type of social interaction (e.g., following the classification by Meier & Reinecke, 2020) and measure corresponding digital skills, as this could provide more precise insight into how people's digital know-how interacts with disconnection practices. Finally, we conducted additional exploratory analyses that revealed insight into the social context in which disconnection takes place. In particular, female participants were less likely to disconnect while being with friends, compared to their male counterparts, which is in line with research showing that phubbing occurs more often among women (Chotpitayasunondh & Douglas, 2016). A potential explanation could be that women generally use digital media more to connect with family and friends (Jackson et al., 2001), making it more challenging for them to disconnect. However, more research into gender roles and related inequalities in disconnection practices is warranted. We also found that older people (age range was from 19 to 75 years) were more likely to disconnect around their partner. It could be that older people are more likely to have a partner, and that this explains our finding. However, it could also be that relationship quality and emotional closeness with one's partner become increasingly important as one ages (Birditt et al., 2009;Carstensen, 1992), resulting in a higher need to disconnect from digital media around one's spouse. Moreover, research on disconnection in older adults has suggested that, unlike younger generations, older adults are less likely to experience conflicts between digital media use and social expectations and responsibilities-and therefore find it less difficult to disconnect (Nguyen et al., 2021). Conversely, we found that younger people were more likely to disconnect around other family members (besides one's children and spouse), friends, and colleagues. A possible explanation for this is that social norms around disconnection are more prevalent among younger generations (Nguyen et al., 2022), and therefore may be more influential on disconnection behaviors among this age group. For future research, it could be fruitful to examine how social norms around disconnection, as well as generational differences in such normative perceptions, relate to the social contexts in which people disconnect. Research on the social context of disconnection may also consider whether people are living alone or are cohabiting with others (e.g., partner and children; Nguyen & Hargittai, 2023). Our exploratory analyses revealed that when people disconnected in the copresence of others (versus when they did so while alone), their subjective well-being improved momentarily. This suggests that the situational, social context in which people disconnect is important for the perceived benefits derived from such intended technology disengagement. This makes sense, as the use of one's smartphone at the expense of focusing on copresent others (i.e., phubbing) is generally found to be disturbing (Vanden Abeele, 2020), which could motivate people to disconnect deliberately during social interactions. Overall, these findings support the hypothesis of "displacement" (Dienlin et al., 2017;Kraut et al., 1998;Valkenburg & Peter, 2007), which implies that the use of digital media may be detrimental to wellbeing when its use goes at the expense of other meaningful (social) activities. The results are also in line with notions from Communicate Bond Belong theory (Hall & Davis, 2017) and related conceptualizations of digital solitude (Campbell & Ross, 2022), which suggest that people invest their social energy in social interactions that benefit their well-being (be it digitally or inperson). In this case, intended disconnection from technology when one is physically co-present could be done to prioritize the in-person social interaction, which might foster connectedness and sense of well-being overall. Conversely, it is also worth considering that some might use digital media as a way of practicing solitude and disconnecting from one's immediate social environment (Keessens, 2023), and, for such people, disconnection then might be less beneficial for well-being. In this study, we only collected data on social context for situations in which people disconnected, so we are unable to test whether individuals who are more often in social contexts and disconnect show different effects compared to those who are more often alone. This could be an important avenue for future research. --- Limitations Several limitations and caveats to this study have been mentioned throughout the article already. Nonetheless, it is important to recap these and note other caveats. First, we used mobile application software to send participants six questionnaires each day over the course of one week. These prompts, as well as reminders when participants did not respond, may have interrupted their moments of disconnection-especially as disconnection is often done from one's mobile phone (Nguyen, 2021). In studying explicit disconnection, other methods of momentary assessments that do not involve people's personal devices, such as a beeper with paper and pencil questionnaires or a separate device just for study purposes, might be preferable. It is important to note that the experience sampling software we used was only available to Android users. Since iOS users tend to have higher incomes, the sample may be biased toward lower income users (Slickdeals, 2018). Furthermore, the use of an experience sampling app (from installation to responding to questions on it) means that participants needed a fair bit of digital skills to make their participation work, which could lead to potential biases as well. This is reflected in participants' relatively high average level of skill in this study, even though we had lower-skilled participants as well. Our study was conducted in The Netherlands, which is a relatively high-income nation with a good digital infrastructure. As such, it is important to realize that the findings might not generalize to other geographic and cultural contexts. Recent research has pointed to digital inequalities in disconnection practices and has called for more scholarship in countries that are less often the focus of inquiry (Bozan & Trere <unk>, 2023;Trere <unk>, 2021). We encourage future research to do so, as well, and to embark on cross-national studies to put findings into a comparative perspective. --- Conclusion On the whole, this study adds much needed empirical evidence to important debates about how the deliberate non-use of digital media relates to subjective well-being. As we find no main effects of everyday disconnection practices on subjective well-being, our research suggests that this relationship might be more complex than previous work has been able to consider. Additional exploratory analyses reveal that people vary greatly in the effect disconnection has on their well-being, ranging from negative to positive outcomes. Moreover, disconnecting when others are physically copresent related to greater well-being as opposed to when people disconnected while they were alone, suggesting that social context plays an important role in how people experience disconnection. Given the complexity of the relationship, whether disconnection practices affect individual well-being may be dependent on moderating factors such as people's motivations for and self-control in disconnecting. While ESM studies are complex and require more resources than many other methods, they are able to capture people's everyday disconnection experiences in their natural setting in ways that are difficult with other approaches. To this end, it is important for scholars to keep innovating methodologically in the realm of disconnection research and continue collecting such nuanced measures in the future. --- Data availability statement The data and analysis code are available on the Open Science Framework (https://osf.io/9chb6/). --- Conflicts of interest statement The authors declare that there is no conflict of interest. --- Open science framework badges Open Data Digitally shareable data necessary to reproduce the reported results are publicly available for this article. --- Preregistered Research design was preregistered.

Drawing on theories of digital media (non-)use and well-being, this study examines how voluntary disconnection relates to subjective well-being and what role digital skills play in this relationship. We rely on mobile experience sampling methods to link nuanced disconnection practices throughout the day (e.g., putting screen devices away and muting notifications) with momentary experiences of well-being. We collected 4,028 responses from 105 mobile media users over the course of one week. Multilevel regression analyses revealed neither significant within-person effects of disconnection on affective well-being, social connectedness, or life satisfaction, nor a significant moderation effect of digital skills. Exploratory analyses, however, show that effects of disconnection on well-being vary greatly across participants, and that effects are dependent on whether one disconnects in the physical copresence of others. Our study offers a refined perspective on the consequences, or lack thereof, of deliberate non-use of technology in the digital age.

Introduction Urbanization in developing countries has been associated with economic development, employment opportunities, improvement in infrastructure and access to better health facilities, all of which are thought to lead to improved health outcomes. [1][2][3][4] On the other hand, rapid, unplanned and unregulated urbanization in developing countries is also associated with a range of factors that are harmful to health and mental health including environmental hazards, such as air pollution and water pollution, poor quality housing, lack of sanitation, road traffic injuries and traffic noise. 2,[5][6][7][8][9][10][11][12][13] In addition, urbanization is also associated with increasing consumption of processed and energy-rich foods and a sedentary lifestyle; these factors can lead to obesity and therefore higher risks of suffering from diabetes, hypertension and heart disease. 10,[14][15][16][17][18][19][20][21] Thus, theories on the impact of urbanization on health, and healthrelated outcomes, in developing countries are controversial; findings vary according to the particular context and the health outcome considered. Urbanization has occurred very rapidly in China. Following market liberalization, the urban population of China has increased from 17.9% in 1978 to 51.3% in 2011, and the number of cities with a population of at least 500 000 has increased from 40 to 140 over this period, 22 with further projected increases under current government policies. 23 Urbanization in China is almost entirely managed by the state, unlike urbanization in other countries. As a result, this process has not created the urban slums which are common features of urbanization in many countries 24 such as India. 25 Urbanization in China can also include forced urbanization, involving involuntary relocation from sub-standard rural dwellings to higherstandard housing in newly created towns. 26 Forced urbanization is associated with a set of positive changes that flow from better housing quality, but it is also thought to relate to problems including disrupted social networks and psychological stress associated with city life, which may bring negative consequences for health and well-being. 27 Although urbanization in China is associated with a number of public health challenges noted above, 4,24,[28][29][30][31][32][33] in general those who move to urban areas have better health outcomes than the populations they leave behind. 34,35 However, those who move are a selective sample in terms of their own health, wealth and other unobserved characteristics that may protect against experiencing adverse health outcomes in the future. [34][35][36][37] It remains unclear then, whether the extent to which the health advantage observed in urban areas is due to contextual characteristics of cities or simply a consequence of health-selective migration. The speed, scale and planned nature of urbanization in China offers a valuable opportunity to investigate the impacts of urbanization on health. We do this by using a nationally representative dataset of older Chinese people to conduct a comparative study of the health outcomes of the in situ, involuntarily urbanized population, i.e. those who live in previously rural areas that have been absorbed by expanding urban areas. This in situ urbanized-rural population group is distinct from those who have always lived in rural or urban locations for the whole of their lives, and the nature and scale of urbanization in China makes this group large enough to study as an independent subgroup of the population. Furthermore, given the pace of urban change in China, the urbanized group of people share early life experience with the rural population and later life experiences with urban dwellers. By studying the differences between these groups, we isolate the effect of urbanization on health and well-being independent of selective migration to urban areas. Under the hypothesis that cities are good for health and that early life advantage predicts late life health, this would lead to four empirical predictions which we explore in this paper. Hypothesis 1. The health and wellbeing of the in situ urbanized population will be (a) better than that of the rural group, and (b) worse than that of the urban --- Key Messages • Urbanization in China is associated with improved self-reported health and lower levels of depression. • This impact of urbanization is independent of the impact of migration that has been noted in the literature in China and elsewhere. • The results add to a literature that suggests that planned urbanization could benefit population health and well-being in countries such as China. • It is likely that improved infrastructure in urban areas is a key driver of the observed health advantages for in situ urbanized rural residents. • However, residents in urban areas also experience higher levels of obesity. group-perhaps with the exception of cardiovascular or chronic diseases where we might expect the effect to be reversed due to lack of exercise and obesogenic diets in urban areas. Hypothesis 2. The health advantage compared with the rural population will not be a consequence of characteristics that could be related to selection. Hypothesis 3. Individual economic mobility, occurring as a consequence of urbanization, will be part of the explanation. Hypothesis 4. The development of urban infrastructure will contribute to this advantage. --- Methods This paper uses the China Health and Retirement Longitudinal Study (CHARLS), a nationally representative dataset that measures the health and well-being of the Chinese population aged 45 and over. 38 The nature of CHARLS is multidisciplinary; it contains detailed information on respondents' social, economic and health conditions. Further details on the sample are provided elsewhere. 39 This paper uses the CHARLS national baseline survey which was conducted between June 2011 and March 2012, and comprises information on about 17 000 individuals and 10 000 households. Our reasons for choosing the CHARLS baseline survey are: first, the CHARLS sample of older adults contains sufficient numbers of people who have been through the urbanization process, which surveys of younger cohorts may not contain; second, the CHARLS includes detailed information on individuals' socioeconomic circumstances and health, including early life circumstances. Precise definitions and measures of urbanization vary, 40 but there is agreement that measures should capture the growing proportion of people who live in cities or urban areas. In this paper, we use the CHARLS classification of an urban area from the National Bureau of Statistics in China, which states that a community is urban if it is located in a city, suburb of a city, a town or other special areas, where non-farming employment constitutes at least 70% of the workforce. Rather than a purely administrative definition, this definition also captures some information on economic activity, which is an important part of the urbanization process. To capture impacts of urbanization, this paper compares in situ urbanized-rural residents with rural and urban residents (N 1<unk>4 12 916). To do this, we first remove migrants and return-migrants from the sample, using data on their current place of residence, birthplace and lifetime migration records of longer than 6 months. We then use Hukou status to identify in situ urbanized-rural residents. The Hukou system is a unique feature in China which is loosely similar to an internal passport system; there are two types of Hukou, an agricultural type and a nonagricultural type; this classification is based on the rural/ urban classification of a person's birthplace. 41 This allows us to separate out urban non-migrants by their Hukou status, so we have urban non-migrants with an urban Hukou and urban non-migrants with a rural Hukou. As the Hukou type (agricultural/non-agricultural) is a marker relating to birthplace, urban non-migrants with a rural Hukou are likely to be a group of rural people who experienced cities being built around as part of the rapid urbanization in China. The analytical value of the urbanized-rural group is that the impact of urbanization on them is likely to be exogenous to health outcomes. --- Empirical model The outcome variables are self-assessed health status (five categories) and depression scores that are calculated based on the 10-item Center for Epidemiological Studies Depression Scales (CESD)-10 items. 42,43 These health outcomes are useful markers of population health, including in terms of predicting future events such as hospital admission, mortality and various clinical outcomes. [44][45][46][47] Both outcome variables were analysed with ordinary least squares (OLS) regressions, where lower values of the dependent variables indicate of better health. Multinomial logit regressions were used to confirm the robustness of the result from the OLS models predicting self-assessed health status. To address the possibility that urbanization relates to growing obesity in China, we also use waist circumference, an objective measure, as an indicator for obesity. 48 The mean and standard deviations of self-reported health status in our analysis sample are 3.03 and 0.92, respectively. These statistics are 8.60 and 6.39 for depression scores and 84.13 and 12.37 for waist circumference, respectively. This paper uses an empirical model that builds on the literature on social determinants of health in later life. We control for demographic, early life, socioeconomic, psychosocial and behavioural factors that relate to health and well-being in later life. 49 We also add an assessment of physical exercise, to explore the possibility that urbanization may relate to changes to a more sedentary lifestyle. This physical activity variable is only available from a random half of the sample in the CHARLS. Furthermore, to indicate the presence of urban infrastructure, a flushable toilet variable is included to capture whether a sewage system in place. This variable has three categories: no toilets in the household, non-flushable toilets and flushable toilets. For early life factors, in addition to education and first job, we also include lower leg length (knee height), which is used as an objective measure of youth and childhood health and socioeconomic circumstances. 50 Supplementary Table 1, available as Supplementary data at IJE online, shows the means of variables with stratification by residential status. The urban group were less likely to have no education, less likely to be farmers and had greater knee height, Encouragingly for our study design, the rural and urbanized-rural groups were similar on these early life factors. For instance, in terms of the first job, 92% of the rural group were farmers compared with 84% of urbanized-rural group, whereas the proportion for the urban group was 28%. However, there are larger differences between the rural and urbanized-rural groups in terms of current socioeconomic variables. For instance, the average value for household durable wealth is 6470 yuan for the rural group, 9520 yuan for the urbanized rural, and 11 460 for the urban group. In terms of marker of the urbanization effect, more than twice the number of urbanized-rural households have a flushable toilet compared with rural households (47% vs 22%), and this figure is even higher for the urban group at 72%. To address hypothesis 1a and 1 b, we show a model controlling for age and gender. To address hypothesis 2, we additionally adjust for early life factors and the extent to which they contribute to the advantage of urban people. Hypothesis 3 is tested by including individual socioeconomic indicators. At this stage of the model we also control for familial support and smoking and drinking behaviours. To address hypothesis 4, we include the indicator of urban infrastructure, having a flushable toilet in the house. We use a sequence of regressions, gradually adding in each cluster of factors from the empirical model to the previous regression model. We have truncated the sample at age 80 to avoid concerns of acute selective mortality (about 3% of the CHARLS sample are over this age and hence excluded from our analysis). Robust standard errors are used in all regressions to allow for heteroskedasticity in residuals and to ensure that appropriate statistical inferences are drawn. Analyses were conducted using STATA 14 (StataCorp, College Station, TX, USA). Full results and results of multinomial logit regressions are in the Supplementary material, available as Supplementary data at IJE online. --- Results Figuress 1-3 depict the means and 95% confidence intervals of self-assessed health status, depression scores and waist circumference, respectively, for the three groups. The urbanized-rural group has better average health status and depression scores compared with the rural group, although experiencing worse scores compared with the urban group. However, the average waist circumference in the urbanized-rural group is smaller than in the urban group but larger than in the rural group. These results hold for both men and women. An interaction term of gender by residential status is not statistically significant. Table 1 shows the results of OLS models for selfassessed health status. The rural group has worse health status compared with the urbanized-rural group (which takes the reference category) and this differential is consistent across all specifications. Control variables do not meaningfully attenuate this relationship until specification 5 of the model. For instance, the coefficient for the rural group is 0.165 (P <unk> 0.001) after adjusting for age and sex in specification 1, and becomes 0.158 (P <unk> 0.001) after controlling for early life factors in step two. In subsequent models, the coefficient stays around this level across all the later model specifications. The rural versus urbanized-rural differential holds even after controlling for the physical exercise variable in step 4, which is based on a half-random sample. The exception to this is after controlling for the flushable toilet in specification 5. This coefficient for the rural group is 0.147 (P <unk> 0.001) in step 4 (controls for demographic, early life and socioeconomic factors) and attenuates by 18% to 0.121 (P <unk> 0.001) in step 5. Furthermore, there is no statistically significant difference in having a non-flushable toilet compared with having no toilet in the household (see the full table in the Supplementary material, available as Supplementary data at IJE online). The results of multinomial logit regressions also confirm this consistent health advantage of the urbanized-rural group. The results for depression scores in Table 2 show similar findings: the presence of statistically significant and consistent advantage in lower depression scores for the urbanized-rural group compared with the rural group; and the only control variable that attenuates this relationship is having a flushable toilet in the household. In this table, the coefficient for the rural group is 1.457 (P <unk> 0.001) after controlling for age and sex. After controlling for the differences in early life factors in specification 2, this coefficient attenuates by 10% to 1.307 (P <unk> 0.001). This coefficient largely stays around this level in later specifications, but it drops by almost 30% from step 3 (1.075, P <unk> 0.001) to step 5 (0.75, P <unk> 0.001), which controls for the flushable toilet. Again, it is having the flushable toilet in the household, as a proxy for improved infrastructure in urban areas, that makes the most difference in explaining the advantage of the urbanized-rural group compared with their rural counterparts. The advantage in health outcomes associated with the urbanized-rural group is also clear when compared with the urban group, after adjusting for compositional differences in early life factors. For instance, in Table 2 for the depression models, after adjusting for age and sex, the coefficient for the urban group is -0.798 (P <unk> 0.001) compared with the urbanized-rural group, indicating better mental health for urban dwellers. This coefficient becomes 0.656 (P 1<unk>4 0.018) after adjusting for differences in early life factors in specification 2. After socioeconomic factors are entered into the model in step 3, this coefficient becomes 0.827 (P 1<unk>4 0.003). A similar magnitude of effect persisted throughout the later models in this table, with a coefficient estimate of 0.905 (P 1<unk>4 0.001) in the fully adjusted model. There is little difference in the self-reported health of the urban and urbanized-rural groups in the baseline model. However, a modest health advantage emerges once early life circumstances are accounted for. Table 3 presents results on waist circumference. After adjusting for age and sex in step 1, the average waist circumference of the rural group is 2.643 cm (P <unk> 0.001) smaller than urbanized-rural group. This difference changes slightly after adjusting for early life factors. After adjusting for socioeconomic factors, this coefficient reduces by a quarter, indicating that this differential relates partly to socioeconomic advantage in urban areas. Further effects are seen when accounting for differences in exercise in step 4, where the coefficient increases in absolute magnitude to -2.331 (P <unk> 0.001), and then for the flushable toilet in step 5. Finally, the urban group has a bigger waist circumference compared with the urbanized-rural group-the coefficient is 2.194 cm (P <unk> 0.001)-when controlling for demographic factors, but this difference is explained away after adjusting for differences in early life and socioeconomic factors. --- Discussion The literature and theoretical perspectives on urbanization and health in developing countries are mixed. 1,2,5,17 In a context of rapid global urbanization, this is an important area of uncertainty. On the one hand the process of urbanization is thought to be associated with improvements in infrastructure and socioeconomic circumstances that lead to improved health outcomes. Yet, on the other hand, scholars point to poorer health outcomes that may be linked to environmental pollution and unhealthy lifestyle choices. This paper investigated the impact of urbanization on health and well-being in China, by building an empirical model to examine health differences and possible causal pathways, developed from the literature on the social determinants of health in later life. A key methodological issue in much of the research on urbanization and health is that the improved health outcomes in urban areas compared with rural regions might, at least in part, be a function of large-scale health-selective migration to urban areas rather than urbanization itself; the characteristics that are conducive to migration from rural to urban areas are related to improved health across the life course. One key strength of this paper is that we were able to identify a population in China that experienced urbanization in situ, and compare them with those who remained in rural areas and those who remained in urban areas throughout their lives, and can thereby investigate impacts of urbanization independent of the influence of migration on health which have also been occurring over this period in China. We found a positive effect of urbanization on health and well-being in China-the urbanized-rural group have improved general and mental health outcomes compared with the rural group. Controlling for early life factors and socioeconomic factors attenuates this relationship only slightly, but controlling for the presence of a flushable toilet largely explained this effect. This variable reflects the construction of a sewage system, which may suggest that the effect is a consequence of a planned urbanization process and associated improved infrastructure. When comparing the urban and urbanized-rural groups, a health/ well-being advantage emerges for urbanized-rural residents after accounting for early life circumstances. In the Chinese context, these urban health/well-being advantages seem to be sufficient to overcome the disadvantages of early life that are more concentrated among those who grew up in rural areas. As other studies have suggested, 24,29 urbanization in China appears to be associated with an increased risk of obesity, as the urbanized-rural group have larger average waist circumferences compared with the rural group. This may be due to the changes in lifestyles, e.g. less physical exercise and more obesogenic diets in urban areas. 24 The trend towards increased obesity in urban areas, and the emergence of dietary risk factors as a leading risk factor accounting for disability-adjusted life-years and deaths 4 in China, may alter urban/rural health differentials in the future. 51 Other studies have directly tested the impacts of urbanization on health in China. 29,51,52 These have measured urbanization effects using an index that includes factors such as community characteristics. The potential problem with this approach is that a range of different mechanisms may be at work in the relationship between these underlying community characteristics and health. Thus, the particular mechanism through which urbanization affects health is unclear from these studies. Additionally, and perhaps more importantly, these studies have failed to address migration processes, which account for a large share of urban population growth in China; thus their results may be biased by factors related to selection into a migrant population. A number of limitations in our analysis should be acknowledged alongside our interpretation of the results. First, the urbanization effect here may also be due to changes in income and the improvement in the health care system in China, both of which will relate to access to urban infrastructure. We do not directly test these here. Additionally, in this paper we are unable to deal with the survival effect, whereby healthier people tend to survive longer than unhealthy individuals, and so we may have underestimated the differences in health status and depression scores and overestimated the differences in obesity. Finally, in this paper we focus on broad measures of health and well-being. Other specific health outcomes, such as conditions affected by air pollution that is a greater issue in cities, might be negatively related to urbanization in China. 53 To conclude, our findings suggest that urbanization in China has, on average, had an independent and positive effect on health outcomes, separate from the potential effect of health selective migration previously noted in the literature. The urbanized-rural and rural groups in our sample have very similar early life circumstances, giving confidence that our comparison of health across these groups is isolating the impact of urbanization on health. The results show that effectively managed urbanization could benefit people's health and well-being, and that improvement in infrastructure is likely to be a key driver. 54 --- Supplementary Data Supplementary data are available at IJE online. Conflict of interest: None declared.

Background: Urbanization in developing countries is usually accompanied by migration to cities, making it a challenge to unpack the independent relationships between migration, urbanization and health, particularly in the presence of health-selective migration. Since 1978, unprecedented planned urbanization has taken place in China and further increases to the urban population are expected. This paper explored the impacts of urbanization in China through a comparative study of in situ urbanized population. Methods: Using the China Health and Retirement Longitudinal Study (CHARLS), a nationally representative dataset for people aged 45 years or older, we compared selfassessed general health, depressive symptoms and waist circumference among three groups: (i) in situ urbanized-rural residents; (ii) rural residents; and (iii) urban residents. Using a model informed by the literature on the social determinants of health in later life, we investigated the patterning and drivers of differences in health outcomes between these three groups, in order to explore the impact of urbanization independent of the impact of migration. Results: There are consistent advantages in health and less depression among urbanized-rural residents compared with the rural group; and this group has even better health outcomes than the urban group after adjusting for early life differences. However, this relationship is reversed for waist circumference. Socioeconomic circumstances and factors related to a planned urbanization partly explain these effects. Conclusions: Urbanization in China has, on average, had an independent and positive effect on health and well-being. Planned urbanization could benefit people's health in developing countries. It is likely that improved infrastructure is a key driver.

Introduction With more than 55% of the world's population living in cities (UN 2018), it is of paramount importance to understand how cities react to and recover from the COVID-19 crisis. In this unprecedented pandemic, various mobility restriction policies have been implemented to curb virus spread, including stay-at-home orders (Nivette et al. 2021) and banning certain indoor businessess 1. These restrictions cause abrupt mobility reduction which poses serious challenges to the operation of urban systems (Tirachini and Cats 2020;Lutu et al. 2020) and even spills impact on online social network usage (Li et al. 2021b). More importantly, as Jane Jacobs famously argued that urban mobility is the underlying social fiber in cities that holds community social structures together through "cross-use of space" (Jacobs 1961), it is crucial to understand how mobility reduction affects social segregation at urban venues. While considerable research attention has been drawn, most previous studies used survey questionnaires or field experiments to examine urban mobility, including park access (Zhang et al. 2022), restaurant patronage (Palacios et al. 2022) and teleworking (Conway et al. 2020). These studies are often limited by the scalability of research methods and only focus on a specific aspect of urban mobility. As online service data are becoming increasingly available as a tool to facilitate understanding of real-world social phenomena (Zhang, Lin, and Pelechrinis 2016;Zhang and Pelechrinis 2014;Cranshaw et al. 2012), several studies leveraged social media check-ins (Han et al. 2021), bikesharing data (Li et al. 2021a) and mobile phone data (Nouvellet et al. 2021) to sense mobility changes at scale, but they zoned in on the response to the initial lockdown in 2020, overlooking the long-term behavior changes in the following years, especially when vaccines have been widely administered and daily life is getting back on track. More importantly, they failed to interpret the implications of mobility change for the social dynamics in urban space. Different from previous research, this study aims to derive a comprehensive picture of urban mobility change at scale across multiple phases of the COVID-19 pandemic, through the lens of the following three research questions: RQ1: Can pervasively collected location service data reveal meaningful patterns of mobility behavior changes during the COVID-19 crisis? RQ2: How do the mobility behavior changes affect the social inclusiveness in urban space? RQ3: Are the mobility behavior changes consistent across geographical regions? To answer these questions, we propose a data-driven analytic framework with the following key designs. First, we leverage a large-scale location service dataset that records 35 billion visits to 2.08 million urban venues in the United States from March 2019 to February 2022. The dataset enables longitudinal analysis to systematically compare the mobility behavior in pre-pandemic, lockdown and reopening periods across all 50 states. Second, we derive a conceptual map based on the relative visitation drop during the lockdown period and the relative visitation recovery in the reopening period to reveal the universal patterns among heterogeneous urban venues. Third, we design a visitation segregation index (VSI) that quantifies the experienced segregation in a specific urban venue as the likelihood of visitors coming from similar socio-demographic backgrounds, and evaluate its association with the change in mobility behavior. Finally, we investigate the geographic differences in mobility change patterns with clustering analyses, and further uncover the links between mobility behavior change and local governing policies. Our analysis under the proposed framework yields several insightful observations. First, most urban venues exhibit a universal pattern, i.e., the relative mobility level at the reopening period scales linearly with the relative mobility level at the lockdown period, which indicates that urban venues suffering from large mobility drops are likely to experience persistent avoidance even after reopening. Second, closer inspection reveals two dominant types, i.e., daily liferelated and work-related patterns, while venues belonging to Parks, Information and Educational categories form three outliers due mainly to their unique urban functions. Third, we find a positive correlation between the experienced income/racial segregation and the level of mobility reduction, indicating that urban venues enduring more mobility drop during lockdowns are more susceptible to social segregation. Finally, the differences in urban mobility change can be largely explained by the political inclination of the residential states. Specifically, the two main clusters of mobility change patterns can be predicted with 81.25% accuracy by partisan inclination, where the Republican-leaning states generally have a higher relative mobility level during both lockdown and reopening periods. In summary, we make the following contributions: 1. We propose a data-driven analytic framework to comprehensively understand the impact of COVID-19 on urban mobility patterns at scale, emphasizing the huge potential of exploring the cyber-physical nexus to investigate real-world social phenomena. 2. We establish a conceptual map to discern between heterogeneous patterns of mobility change associated with different types of urban venues, and dig into the underlying reasons regarding distinct urban functions. 3. We find a positive correlation between the experienced segregation and mobility reduction at urban venues, which suggests maintaining urban mobility is indeed important to healthy neighborhood social fabrics. 4. We show that the main patterns in state-level urban mobility change can be accurately predicted by partisan inclination, which provides a potential mechanism to explain the different responses between states. --- Related Work We categorize related works along three dimensions. Evaluating the impact of COVID-19 on urban venue accessibility. A variety of works have provided insights on how the COVID-19 pandemic impacts urban systems, especially the accessibility of urban venues, which can be classified into two categories based on the research method. (1) Survey-based. Zhang et al. (2022) conduct a questionnaire survey on the changes in park visits change, and find that higher epidemic risk leads to lower park accessibility and greater inequality in park access. Conway et al. (2020) conduct a survey among highly-educated US residents on the impacts of COVID-19 on work and life patterns, and discover a decreasing demand in restaurant patronage and an increasing hope to continue teleworking after the pandemic. Albeit offering valuable insights, survey-based studies are inevitably influenced by self-report bias and constrained by sample sizes, and thus unable to provide a comprehensive picture of the mobility changes. Different from them, we base our analysis on large-scale mobility data collected from online location services, which contains rich and finegrained information about people's real-world movements. (2) Sensing-based. Han et al. (2021) utilize check-in data to identify several "early bird" categories that exhibit visit drops before the city lockdown. Li et al. (2021a) use bikesharing data to study inter-city mobility during lockdowns and find a decrease in the share of workplace visits but an increase in that of park and grocery store visits. These works overcome the limitation of self-report bias, but did not provide a systematic analysis of mobility changes across different phases of the pandemic. Different from them, we jointly analyze the mobility patterns in three phases of the COVID-19 pandemic, and conceptualize a systematic framework to classify various urban venues by mobility change patterns. Leveraging online location service data to understand real-world social phenomena. Digital traces collected from online service platforms, including check-in, geo-tagging and passive geo-positioning data, provide fine-grained information about human activities (Cranshaw et al. 2010;Cho, Myers, and Leskovec 2011), which have huge potential to uncover the principles governing physical activity patterns. Using such data, researchers discovered latent structures in urban transition flows (Zhang, Lin, and Pelechrinis 2016), revealed the mechanisms behind spatial homophily (Zhang and Pelechrinis 2014), and renovated notions of activity areas beyond traditional municipal boundaries (Cranshaw et al. 2012). This line of works derives critical insights into the emergence of collective behavioral patterns from individual movements. Complementary to these works, we provide insights into the change of human mobility patterns during different phases of a pandemic crisis. Online service data also shed light on solving various real-world social problems, e.g., estimating urban inequality (Ganter, Toetzke, and Feuerriegel 2022), predicting crime concentrations (Kadar et al. 2020;Rumi, Shao, and Salim 2020), and tracking public concerns (Wang and Taylor 2018). In this work, we reveal the impact of COVID-19 on experienced segregation which is an important aspect of neighborhood social fabrics. Urban mobility and experienced segregation. A long history of research has been focused on residential segregation in urban spaces, where people of the same racial/ethnic backgrounds and similar socioeconomic status tend to reside closer to each other and thus form distinct neighborhoods (Charles 2003;Taeuber and Taeuber 2008;Lichter, Parisi, and De Valk 2016). However, such static segregation metrics are insufficient in reflecting people's daily experience of social isolation when traveling in urban environments. Thus, some recent studies propose a type of de facto segregation emerging from people's everyday mobility behavior, referred to as experienced segregation. Wang et al. (2018) utilize social media data to exhibit the persistent isolation between neighborhoods of different socioeconomic statuses despite long travel distances. Moro et al. (2021) use mobile phone positioning data to further reveal the association between individual visitation preferences and the experienced segregation in different kinds of places. However, existing works study the relationship between mobility and segregation in a single snapshot, without considering how changes in the former may bring about changes in the latter. Complementary to them, we use different phases of the COVID-19 pandemic to form a natural experiment, so that we can observe the changes in experienced segregation when substantial changes are imposed on urban mobility patterns. Urban mobility and partisanship. Classic works on partisanship and mobility typically focus on long-term outcomes such as migration-induced geographical assortativity (Gimpel and Hui 2015;Tam Cho, Gimpel, and Hui 2013), and social mobility (Dancygier and Saunders 2006). More recent works turn to short-term physical mobility, analyzing its connection with people's political inclinations. It is found that partisan differences are associated with people's willingness to reduce mobility and adopt other prevention measures (Clinton et al. 2021;Bruine de Bruin, Saw, and Goldman 2020). Such a relation is mediated by a range of subjective factors, including perceived health risk, perceived effectiveness of prevention measures, and optimism about a forthcoming end of the pandemic (Stroebe et al. 2021;Freira et al. 2021). While these studies base their results on questionnaires for individuals, we provide a holistic view of such behavior gaps throughout multiple pandemic phases by directly looking into human mobility data. --- Data Description In this study, we mainly use three types of data, i.e., POI data, mobility data, and demographic data. For POI data, we leverage Safegraph's Places dataset2, which is collected from open-source web domains and publicly available APIs. It contains detailed information about a POI's name, address, industrial type, etc. For mobility data, we leverage Safegraph's Patterns dataset3, which is a large-scale dataset that records visits to POIs in the US from online applications with location services. It records the number of visits paid to each POI by residents of each CBG, aggregated on a monthly basis, along with supplementary statistics such as popularity by week and by hour. Although this dataset is a sample of all happened activities, existing works (Kang et al. 2020;Brelsford et al. 2022) have validated its consistency with other datasets and representativeness of the real-world situation. To account for uneven sampling across CBGs, we follow the common practice (Chang et al. 2021) to reweight POI visitations according to the ratio between the number of devices residing in each CBG to the corresponding CBG. We select the POIs located in urbanized areas according to the rural-urban classification scheme provided by the U.S. Department of Agriculture4, and further group POIs into industrial categories based on the North American Industry Classification System (NAICS) 5. Specifically, we focus on the following 15 representative POI categories: Parks, Educational, Information, Retail, Health Care, Recreation, Hospitality, Construction, Manufacturing, Wholesale, Transportation, Finance, Professional, Community Services, and Public Admin. We summarize the semantic meanings and statistics of POI categories in Table 1. The monthly visitation frequencies are shown in Figure 1. We observe that the overall visitation frequency dropped significantly after Feb. 2020, when the COVID-19 pandemic started to surge in the US. After remaining low for months, the overall visitation frequency began to rebound and has remained on a higher level since May 2021, which is consistent with the timeline of reopening6. Based on these observations, we take mobility data from three distinct phases of the pandemic for comparative analysis. The pre-pandemic phase is represented by Jan. 2020 -Feb. 2020, before the WHO declaration of COVID-19 as a pandemic (Cucinotta and Vanelli 2020). The lockdown phase is represented by Jan. 2021 -Feb. 2021, when national and local authorities put into effect a series of COVID-related restriction policies in face of the surging numbers of COVID-19 cases and deaths. The reopening phase is represented by Jan. 2022 -Feb. 2022, when most parts of the country have lifted restrictions and gone back to business as usual7. To link mobility behavior to visitor attributes, we obtain demographic data from the 2019 American Community Survey (ACS) 5-year Estimates, which record demographic compositions of census block groups (CBGs). As the smallest geographical unit with publicly-released demographic statistics and a typical population between 600 and 3,000, CBGs provide a suitable scale for analyzing socioeconomic differences. We divide CBGs into four income quartiles according to median household income, and four race quartiles according to the percentage of residents belonging to racial/ethnic minorities (Van Voorhees et al. 2007), and use them to quantify experienced segregation in venue visitation. --- Methods Profiling Urban Mobility Behavior Changes We denote the visitation frequency in the pre-pandemic, lockdown, and reopening phases as F p, F l, and F r, respectively. For each POI category, we depict the changes in the associated visitation patterns with two metrics. First, we calculate the relative mobility level during COVID-19 lockdowns (denoted as M l ) as the change of visitation frequencies with respect to the pre-pandemic level: M l = F l -F p F p.(1) Second, we calculate the relative mobility level after reopening (denoted as M r ) as the change of visitation frequencies also with respect to the pre-pandemic phase: M r = F r -F p F p.(2) Taking these two metrics as Euclidean coordinates, we construct a two-dimensional conceptual map of POI categories, where proximity indicates similar patterns of mobility drop and recovery. We will elaborate on the conceptual map and the classification results in the Results section. --- Measuring the Impact on Experienced Segregation Seemingly free to move around in urban spaces, people from "mainstream" neighborhoods and disadvantaged neighborhoods still demonstrate considerably different patterns in venue visitation, due to both personal choices and capability to access important urban facilities. Such experienced segregation is ubiquitous, dynamic, and thus serves as a timely indicator of changes in individuals' social preferences, lifestyles, and the society's inclusiveness. To quantify the changes in experienced segregation throughout different phases of the COVID-19 pandemic, we devise a visitation segregation index (VSI) from the framework proposed by Moro et al. (2021). First, we divide CBGs into quartiles according to their different demographic characteristics, and assume people from the same CBG are associated with similar socioeconomic status, which is consistent with existing literature (Chang et al. 2021). As our mobility data records visits paid to each venue by people from each CBG, we aggregate CBGs within the same quartile to calculate the proportion of visitors from each quartile to a specific venue, denoted as p i for the i-th quartile. Finally, we calculate the experienced segregation as the deviation of the realistic visitor composition from the ideal case where people from all quartiles evenly visit this venue. The resulting expression is shown below: V SI = 2 3 4 i=1 p i - 1 4.(3) Note that by adding a constant multiplier to the expression, we ensure that the value falls within [0, 1]. Specifically, if people from all quartiles visit a venue equally, we have p i = i <unk> = x, which makes VSI= 2 3 • |1 -1 4 | + 3 • 2 3 • 0 -1 4 = 1, indicating the highest segregation level. Specifically, we focus on two types of experienced segregation, i.e., income segregation and race segregation. Experienced income segregation is calculated based on population division by median household income, and experienced race segregation is calculated based on population division by the percentage of racial/ethnic minorities. --- Discovering Patterns of Mobility Behavior Changes across Geographical Regions To analyze the geographical patterns of mobility behavior changes in different states, we first characterize each state with a 30-dimensional vector, which is constructed by concatenating the state-level mobility change in the 15 categories during two phases: the lockdown phase and the reopening phase. On the feature vectors, we perform agglomerate clustering, a bottom-up approach that discovers clusters by successively merging data points. Specifically, we set the goal to be minimizing the average Euclidean distances between all data points belonging to any pair of clusters. --- Results --- Heterogeneous Mobility Behavior Changes in Urban Space (RQ1) Taking the mobility level during COVID-19 lockdowns (M l ) and the mobility level after reopening (M r ) as coordinates, we obtain a conceptual map of POI categories, as shown in Fig. 2(a). Most of the categories (12 out of 15) are distributed along a straight line, except three. Each of these three categories can be traced to reasons acknowledged in literature or news report, which we will illustrate later. For the 12 "mainstream" categories, Ordinary Least Squares (OLS) regression identifies this line as M r = M l + 0.077 (RMSE = 0.051, R 2 = 0.808), and all the 12 categories lie in its 99% confidence interval (CI). The high goodness-offit of the regression model validates the linear relationship between M r and M l, indicating a similar mobility elasticity shared by these 12 categories. In other words, the mobility recovery is not proportional to the mobility drop. Instead, venues with a smaller mobility drop can expect a larger relative mobility recovery, while great mobility drops are likely to persist through the whole pandemic. Taking Retail as an example, we observe that its mobility drop in the lockdown phase is very small, and thus its mobility level in the reopening phase has almost returned to the pre-pandemic level. By contrast, Manufacturing endures a large mobility drop, and its mobility in the reopening phase only restores to 40% of its pre-pandemic level. By further analyzing the values along both axes and combining them with the semantics of these categories, we discover that these 12 categories are linearly separable by M r = -0.2 in the two-dimensional space, resulting in two types that correspond to work-related venues and life-related venues, respectively. Specifically, Type 1 consists of 8 categories (Construction, Wholesale, Finance, Public Admin, Professional, Manufacturing, Transportation, and Community Services), all of which see a relatively small mobility recovery. As these categories are mainly working places, the visitations to the corresponding venues decreased due to the widely-adopted remote working policies aiming to reduce transmission risk. Albeit with certain mobility recovery, the magnitude of mobility remains clearly smaller than the prepandemic level, which indicates a likely-perpetual paradigm shift in people's working modes, i.e., from on-site to online. By contrast, Type 2 categories recover better towards the pre-pandemic level. This type consists of 4 categories (Retail, Health Care, Recreation, and Hospitality), which are essential businesses that support citizens' everyday needs. Thus, visitations to such venues display greater capability to resume normal. Finally, each of the three remaining categories exhibits a distinct pattern of mobility change. Type 3 (Information) sees the most remarkable mobility drop and most petite mobility recovery, likely because many high-tech businesses in this category are highly capable of shifting to remote working. Type 4 (Educational) sees both a large mobility drop during the lockdown and a substantial mobility recovery after reopening. The main reason is that educational venues such as K-12 schools and universities are highly compliant with administrative policies, thus explicitly instructed to be closed during the most severe phase of the pandemic and gradually resume normal teaching afterwards8. Finally, Type 5 (Parks) turns out to be the only category that experiences no reduction but an increase in mobility during the lockdown. The main reason is that those who have gone through stay-at-home orders feel a greater need to socialize with other people. Since indoor venues are either closed or risky in transmitting viruses, many people turn to show greater passion for gardens, city parks, and other green open spaces (Curtis et al. 2021;Lu et al. 2021;Venter et al. 2020). Next, we analyze how the proportion of visitations by POI category (Figure 3) changes through different phases of the pandemic. The top-5 categories receiving the most visitation before the pandemic are Retail, Hospitality, Educational, Health Care, and Recreation. During the lockdown phase, the top-4 categories remain the same, despite that Health Care and Educational exchange their positions. Meanwhile, Parks replaces Recreation to become the category receiving the fifth most visitations, consistent with our previous observation. After reopening, the top-5 categories resume the pre-pandemic order, although with different proportions of visitations. The restoration of the top-5 categories exhibits a certain level of resilience in the composition of visitation. The Chi-square independence test shows that the differences between category-wise visitations in different periods are statistically significant (lockdown & pre-pandemic: <unk> 2 = 21651683.3, df = 6, p = 0.00, reopen & pre-pandemic: <unk> 2 = 5867727.8, df = 6, p = 0.00, reopen & lockdown: <unk> 2 = 12943491.3, df = 6, p = 0.00). These categories can be further divided into three tiers (Fig. 4): Retail is the top tier with the largest proportion of visitations; Hospitality, Educational, and Health Care con- stitute the mid tier, and all the other categories make up the bottom tier. Comparing the lockdown phase with the prepandemic phase, we observe an increase in the proportion of visitations to both top-tier (from 29.7% to 34.9%) and bottom-tier (from 19.0% to 19.5%) categories, while midtier categories suffer a shrink (from 51.2% to 45.6%). We make sense of this observation by pointing out the relation between proportion expansion/shrinkage and the needs satisfied by different venues. The top-tier category of venues satisfy the common essential needs of urban residents, and the bottom-tier includes a wide range of services to satisfy diverse, highly-personalized needs. On the contrary, the mid-tier categories are either for satisfying optional needs or required to be closed during lockdowns, and thus the proportion of visits to such categories shrinks. The Chisquare independence test shows that the differences between category-wise visitations in different periods are statistically significant (lockdown & pre-pandemic: <unk> 2 = 12493730.6, df = 4, p = 0.00, reopen & pre-pandemic: <unk> 2 = 3176123.5, df = 4, p = 0.00, reopen & lockdown: <unk> 2 = 11655823.9, df = 4, p = 0.00). Later in the reopening phase, there is again a trend of returning to the pre-pandemic normal, with the proportion of visitations to the mid-tier rising to 50.1%. Our findings are aligned with Maslow's hierarchy of needs (McLeod 2007): during crises, people tend to focus more on fulfilling their basic physiological needs, and suppress higher-level needs including aesthetics, esteem and self-actualization. Such a change is transient and reversible, with higher-level needs naturally reviving as long as people feel their lowerlevel needs satisfied. Third, we analyze how the weekday/weekend popularity of different venues changes with time (Fig. 5). First, certain patterns persist throughout different pandemic phases. For example, Type 2 venues have the greatest popularity on both weekdays and weekends before the pandemic, and the latter is even greater than the former by over 8 percentage points. In the lockdown phase, both popularity increase, making it still the most popular type. Moreover, we find that for different types of POIs and different periods of time, changes in weekday popularity are highly similar to those on weekends. The only exception is Type 2, whose weekday popularity decreases but weekend popularity increases in the reopening phase. This again indicates changes in residents' lifestyles brought by the pandemic. For example, people may now prefer a big purchase on weekends to store enough food and daily necessities, instead of dropping by convenience stores on the way to and back from work on weekdays. We also observe that the weekday and weekend popularity of Type 1 venues keeps decreasing during the lockdown and after reopening. This resonates with our observations from the conceptual map in Fig. 2 to indicate that changes in people's working paradigm reduces the need to visit Type 1 venues, which is likely to persist even after exit from the pandemic. --- Retail --- 3.6% --- Understanding the Implication for Experienced Segregation (RQ2) In this section, we tease out how experienced segregation changes throughout different phases of the COVID-19 pandemic. As a preliminary, Fig. 6 shows different levels of income and race segregation associated with different categories of venues in the pre-pandemic phase. The highest level of segregation is observed in Type 1 venues (from Construction to Public Admin), while the lowest level is observed in Type 2 venues (from Retail to Hospitality). We also observe a clear correlation between experienced income segregation and experienced race segregation, which is reasonable as racial/ethnic minorities are more likely to be disadvantaged in economic status. Aggregating the changes in all POIs, we observe an increase in both income segregation (<unk> 7%) and race segregation (<unk> 5%) in the lockdown phase, alerting that different classes of people become more socially isolated in the pandemic, which may undermine the integration of the society. The trend is reversed to some extent in the reopening phase (<unk> 3% for income segregation and <unk> 2% for race segregation), but both kinds of experienced segregation are still above their pre-pandemic level, indicating the long-term impact of the pandemic. We perform paired samples t-tests to validate the statistical significance of differences. For income segregation, all three pairs of phases show statistical significance (lockdown & per-pandemic: t=-183.4, p=0.00, reopen & pre-pandemic: t=-94,7, p=0.00, reopen &lockdown: t=-95.5, p=0.00). Also, for race segregation, all three pairs of phases show statistical significance (lockdown & per-pandemic: t=-151.7, p=0.00, reopen & pre-pandemic: t=-82.3, p=0.00, reopen &lock-down: t=-74.5, p=0.00). Next, we disaggregate the changes onto five venue types, and find different patterns of segregation change. Fig. 7(a)(c) show that during the lockdown phase, Type 3 (Information) venues see the greatest increase in both types of experienced segregation, followed by Type 4 (Educational), Type 1, Type 3 (Information) and Type 5 (Parks) venues. In the reopening phase, all types of venues except Type 5 (Parks) see a reduction in experienced segregation, with the greatest reduction in Type 3 venues, as shown in Fig. 7(b)(d). Moreover, we discover an underlying connection between the change in experienced segregation and the magnitude of mobility drop/recovery in the corresponding phases. Recall that in the lockdown phase, Type 3, Type 4, Type 1 and Type 2 venues experience mobility drop from largest to smallest, while Type 5 venues see an increase (Fig. 2). For the first four types of venues, the more the mobility drops, the more the experienced segregation will increase. The main reason is that visitations to such venues are largely through "utilitarian" movements, i.e., people travelling purposefully to reach certain working or shopping places (Hunter et al. 2021). Such movements are generally associated with longer travel distances across neighborhoods, which contributes to a more homogeneous population mixing. By contrast, visitations to Type 5 belong to "leisure" movements, e.g., people walking around casually for relax or entertainment, which usually take place near their residential area. Moreover, the pandemic has strengthened this characteristic by increasing people's willingness to visit closer parks (Zhang et al. 2022). Thus, an increase of Type 5 visitations is likely to resonate with the existing residential segregation to strengthen the experienced segregation in Type 5 venues. As for the reopening phase, visitations to the first four types of venues recover to different extents, which corresponds to the reduction in experienced segregation. For Type 5 venues, since visitations keep increasing throughout the three phases, the experienced segregation does not change in a statistically significant manner compared to the lockdown phase. To provide a finer-grained picture, we step down from the type level to the category level. For each pandemic phase and each type of segregation, we perform linear regression to predict the median segregation change of each venue category (y) using the median mobility change of each venue category (x). For the lockdown phase, we got y = -0.1753x -0.0260 (R 2 = 0.5613, p <unk> 0.05) for income segregation, and y = -0.1083x -0.0048 (R 2 = 0.3544, p <unk> 0.05) for race segregation. For the reopening phase, we got y = -0.0905x -0.0152 (R 2 = 0.4314, p <unk> 0.05) for income segregation, and y = -0.0673x -0.0019 (R 2 = 0.3158, p <unk> 0.05) for race segregation. We make two observations from the above regression results. First, albeit simple, such linear regression models can explain a considerable portion of the variability of segregation, indicating that the connection we identify between mobility change and segregation change is still robust on the category level. Second, the regression slope is steeper in the case of experienced income segregation (-0.1753 in the lockdown phase and -0.1083 in the reopening phase), compared to the experience race segregation (-0.0905 in the lockdown phase and -0.0673 in the reopening phase). As income segregation is more sensitive to mobility change, the widening gap between different income groups may raise more concerns during pandemic crises. --- Analyzing the Differences between States (RQ3) In this section, we analyze the geographical patterns of mobility change in 50 U.S. states. By clustering the feature vectors of the states (see the Methods section), we group the 50 states into 2 distinct clusters, leaving three states as "outliers". As shown in Fig. 8, Cluster-1 states are mainly located in the west and the east, while Cluster-2 states are mainly located in the central part of the country, displaying clear geographical patterns. Category-by-category comparison in Fig. 10, 11 shows that the two clusters differ significantly in the magnitude of mobility change: visitations to all categories of venues in Cluster 1 see a greater mobility drop during the lockdown phase, and remain lower in the reopening phase compared with their pre-pandemic level. Moreover, we put the cluster map alongside the political inclination map derived from the 2020 U.S. presidential election statistics (Fig. 9), and observe a clear correspondence in between: Cluster 1 mainly consists of Democratic-leaning states, while Cluster 2 mainly consists of Republican-leaning states. In fact, even if we simply assign Democratic-leaning states to Cluster 1 and Republicanleaning states to Cluster 2, respectively, we are already able to achieve a hit rate of 81.25%. To rule out the confounding effects of income and race segregation, we perform regres-sion analyses to predict state-level mobility change. Specifically, we compare two linear regression models. The first one only takes state-level political inclination as input, while the second one takes as input political inclination, race segregation and income segregation. In both lockdown and reopen phases, it turns out that state-level political inclination on its own already predicts mobility change pretty well (R 2 = 0.557 for lockdown and R 2 = 0.419 for reopen). Moreover, after taking into account income and race segregation, the state-level political inclination is still statistically significant (p = 0.000 for both lockdown and reopen) for the prediction of mobility change. From the above we note that political inclinations are significantly correlated to mobility change even when we control for differences in income and race segregation, thus further consolidate our results. Everlasting debates between the two parties over coronavirusrelated policies since Mar. 2020 can potentially explain the association. Survey-based results (Pickup, Stecula, and Van Der Linden 2020;Clinton et al. 2021;Center 2020) reveal that Democrats are more compliant with general mobility control and adoption of personal preventive measures such as mask-wearing, while Republicans appear more optimistic about the pandemic situation, and thus feel more comfortable in dining out, shopping and attending indoor events9. Such behavioral preferences are captured by Fig. 10, 11 in both the overall tendency of greater mobility reduction in Cluster-1 states and the differences in the relative magnitude of changes in each category with respect to the cluster's average mobility change. In Cluster 1, mobility drops in Transportation, Finance, Professional, Public Admin and Hospitality are substantially greater than that in Cluster 2. Since these categories cover various aspects of urban life from

Understanding the impact of COVID-19 on urban life rhythms is crucial for accelerating the return-to-normal progress and envisioning more resilient and inclusive cities. While previous studies either depended on small-scale surveys or focused on the response to initial lockdowns, this paper uses large-scale location service data to systematically analyze the urban mobility behavior changes across three distinct phases of the pandemic, i.e., pre-pandemic, lockdown, and reopen. Our analyses reveal two typical patterns that govern the mobility behavior changes in most urban venues: daily life-centered urban venues go through smaller mobility drops during the lockdown and more rapid recovery after reopening, while work-centered urban venues suffer from more significant mobility drops that are likely to persist even after reopening. Such mobility behavior changes exert deeper impacts on the underlying social fabric, where the level of mobility reduction is positively correlated with the experienced segregation at that urban venue. Therefore, urban venues undergoing more mobility reduction are also more filled with people from homogeneous socio-demographic backgrounds. Moreover, mobility behavior changes display significant heterogeneity across geographical regions, which can be largely explained by the partisan inclination at the state level. Our study shows the vast potential of location service data in deriving a timely and comprehensive understanding of the social dynamic in urban space, which is valuable for informing the gradual transition back to the normal lifestyle in a "postpandemic era".

(p = 0.000 for both lockdown and reopen) for the prediction of mobility change. From the above we note that political inclinations are significantly correlated to mobility change even when we control for differences in income and race segregation, thus further consolidate our results. Everlasting debates between the two parties over coronavirusrelated policies since Mar. 2020 can potentially explain the association. Survey-based results (Pickup, Stecula, and Van Der Linden 2020;Clinton et al. 2021;Center 2020) reveal that Democrats are more compliant with general mobility control and adoption of personal preventive measures such as mask-wearing, while Republicans appear more optimistic about the pandemic situation, and thus feel more comfortable in dining out, shopping and attending indoor events9. Such behavioral preferences are captured by Fig. 10, 11 in both the overall tendency of greater mobility reduction in Cluster-1 states and the differences in the relative magnitude of changes in each category with respect to the cluster's average mobility change. In Cluster 1, mobility drops in Transportation, Finance, Professional, Public Admin and Hospitality are substantially greater than that in Cluster 2. Since these categories cover various aspects of urban life from public service to business to personal consumption, the result indicates that partisan gaps span pervasively into all walks of life, acting as a strong power that regulates the social behavior of urban residents. Our observations corroborate with previous research findings that both COVID-19 policy stringency and citizen responses are substantially influenced by political inclination (Adolph et al. 2022;Clinton et al. 2021;Center 2020). While previous studies either use issued mandates or conduct surveys, our analysis from large-scale online service data provides a finer-grained and more realistic picture. Moreover, our results provide evidence of people's differentiated actions beyond their differentiated attitudes toward restriction policies. Nevertheless, we note that our analysis method does not indicate causality or temporal order, thus readers should be cautious in interpreting our findings in a causal way. --- Discussion Our findings provide valuable insights for researchers, urban planners and business owners. For researchers, we provide a universal and easily extendable analytic framework of mobility behavior change, which can be adopted to study similar problems on different spatial or temporal scales. Our concrete findings drawn from passively-sensed large-scale data enhance previous studies with a comprehensive and realistic view, and deepen the understanding of urban mobility and its relationship with other important characteristics in urban life. Our analysis also evidences the value of online location service data in understanding real-world social phenomena, which entail much richer information about human-human relationships and human-location interaction that transcends traditional survey data. For urban planners, the five distinct patterns of visitation change we identify can be used as a reference when devising targeted policies to ease the burden of certain industrial sectors during epidemics. For example, big tech companies are usually more comfortable in shifting working modes to quickly adapt to the changing environment, but traditional industries or small businesses may suffer substantial pain in the transition, and thus the latter should receive more guidance in risk management and more support during hard times. The heterogeneity of recovery capability also highlights the importance of urban resilience, which can be improved preemptively by accelerating technical revolutions of traditional industries and maintaining a proper composition of different industrial sectors. For business owners, the mobility change patterns we find serve as a "teaser" about the business responses during crises and inform better preparations for future disturbance. For instance, it is important for business setups falling in vulnerable categories to improve their risk management and establish potential pathways to transform their business models (e.g., online order and delivery) if the current model is disrupted. Our work has several limitations. First, the location service data used in this study have been aggregated to the monthly level, constraining our capability to analyze more granular temporal changes. If finer-grained data are available, we can naturally extend the current analysis to understand more subtle changes in mobility rhythms. Second, we classify urban venues at a relatively high level and may overlook certain intra-category differences. Nevertheless, our framework for estimating changes in urban mobility and ex-perienced segregation can be readily applied to investigate into any specific type of urban venues. Third, our analysis mainly reveals correlations rather than causalities between features. As future work, we will explore causal analysis to get rid of the influences of confounders. --- Conclusion In this work, we utilize a large-scale location service dataset to systematically analyze the changes in urban venue visitation patterns in three phases of the COVID-19 pandemic. We conceptualize five patterns of mobility change, and rationalize the patterns with the nature of different venues and their function in the hierarchy of human needs. We analyze the change in experienced income and race segregation in different venues, and establish the link between mobility drop/recovery and change of experienced segregation. We discover two clusters of states with similar patterns of mobility change, and demonstrate the high resonance between their geographical distribution and the distribution of Democrat-leaning/Republican-leaning states. Our work provides an extendable framework for jointly analyzing urban mobility patterns across multiple time periods, and our findings contribute to better planning of urban resource distribution in the face of crises. --- Ethics Statement The mobility data from Safegraph are aggregated to the CBG level on a monthly basis, and thus do not contain any individual-level data. To further enhance privacy, differential privacy techniques (https://docs.safegraph.com/docs/ monthly-patterns<unk>#privacy) are applied, and groups with too few visitors are removed. The demographic data from American Community Survey are publicly available at https:// www.census.gov/programs-surveys/acs/, which also reports data on the CBG level. Therefore, no approval from the Institutional Review Board was required by the authors' institutions.

Understanding the impact of COVID-19 on urban life rhythms is crucial for accelerating the return-to-normal progress and envisioning more resilient and inclusive cities. While previous studies either depended on small-scale surveys or focused on the response to initial lockdowns, this paper uses large-scale location service data to systematically analyze the urban mobility behavior changes across three distinct phases of the pandemic, i.e., pre-pandemic, lockdown, and reopen. Our analyses reveal two typical patterns that govern the mobility behavior changes in most urban venues: daily life-centered urban venues go through smaller mobility drops during the lockdown and more rapid recovery after reopening, while work-centered urban venues suffer from more significant mobility drops that are likely to persist even after reopening. Such mobility behavior changes exert deeper impacts on the underlying social fabric, where the level of mobility reduction is positively correlated with the experienced segregation at that urban venue. Therefore, urban venues undergoing more mobility reduction are also more filled with people from homogeneous socio-demographic backgrounds. Moreover, mobility behavior changes display significant heterogeneity across geographical regions, which can be largely explained by the partisan inclination at the state level. Our study shows the vast potential of location service data in deriving a timely and comprehensive understanding of the social dynamic in urban space, which is valuable for informing the gradual transition back to the normal lifestyle in a "postpandemic era".

Introduction As elsewhere around the globe, the spread of the coronavirus disease 2019 (COVID-19) posed formidable challenges to healthcare systems in the Middle East and North Africa (MENA) region. With the first case confirmed in February 2020, the virus rapidly spread through countries in the region (that is, Algeria, Bahrain, Egypt, Iran, Iraq, Israel, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Palestine, Qatar Saudi Arabia, Syria, Tunisia, United Arab Emirates and Yemen), causing more than 224,000 confirmed cases and nearly 8,400 deaths by May 2020 [1]. As at February 2023, the WHO Coronavirus Dashboard shows that MENA countries have reported more than 26 million confirmed cases and over 323,000 deaths [1]. Generally, the severe spread of the virus in MENA countries has been attributed to the influence of a suite of institutional and contextual factors, including the healthcare systems' lack of preparedness, the inadequacy of health resources and management strategies, governments' denial of the seriousness of infection among the population in the early stages of the pandemic, and widespread conflict and migration in the region that hampered individuals' access to healthcare [2,3]. In addition, evidence has accumulated to show that the public's non-compliance with MENA governments' preventive measures (e.g. social distancing, mandatory use of face masks, and repeated handwashing) contributed to the spread of the virus in MENA countries, as well as to the emergence of severer subsequent waves of infection in late 2020 and 2021. This turned many countries into epicentres for viral spread, leading to high mortality rates and damaging social and economic repercussions [4][5][6][7]. From a health-behaviour theory perspective, risk perception, defined as an individual's perceived susceptibility to a threat, is a key factor in shaping individuals' health behaviour and a predictor of such behaviours being maintained [8,9]. In other words, the extent to which individuals believe that COVID-19 infection is severe and likely to affect their health should predict their intention and actual behaviours in relation to compliance with risk prevention and mitigation measures, such as using a face mask, maintaining social distancing and frequent handwashing [9]. Therefore, developing a deeper understanding of how the perceived severity of and susceptibility to threats influence people's behavioural responses and compliance with governments' mitigation measures during times of crises (e.g. the COVID-19 pandemic) is crucial in order to formulate effective public health strategies for risk prevention and mitigation [10]. Building on the Health Belief Model (HBM) developed by Rosenstock [11], and validated in a large body of empirical studies [12], the present study investigated the association between individuals' concern about contracting COVID-19 and their engagement in three preventive behaviours, namely wearing face masks, maintaining social distancing and handwashing. The study used panel dataset from five MENA countries (Egypt, Jordan, Morocco, Sudan, and Tunisia) with 34,219 observations, and employed a probit estimation technique to empirically examine this association, while controlling for country and survey wave fixed effects, sociodemographic characteristics (gender, age, income and education level) and community characteristics (rural versus urban areas). Specifically, the study addressed three important limitations in existing literature on the association between risk perception as behavioural triggers and preventive behaviours in the context of the COVID-19 pandemic. First, despite the extensive literature published on risk perception [13,14] and preventive behaviours [15,16] individually, very few studies consider the association between them. In connection with this, the level of an individual's concern about infection has received very little attention in this literature [17] compared with other behavioural triggers such as knowledge and attitudes [18] or psychological disorders such as anxiety and depression [19], despite the crucial role played by individuals' level of concern about infection in shaping their health behaviours [20]. Thus there is a research gap with regard to understanding the influence of people's level of perception on their subsequent preventive behaviours. Second, most of the earlier studies have focused on a single wave of the pandemic, neglecting the evidence showing that individuals' perception and concern about COVID-19 infection, and thus their protective behaviour, underwent substantial changes over the course of the pandemic and across various waves [21]. Such changes largely depended on infection rates, media coverage and public policy and communication in relation to the implementation of migration measures [22]. Our study used multiple waves of the Combined COVID-19 MENA Monitor Household survey spanning from November 2020 to August 2021, and controlled for countryspecific and time trends. This enabled changes within a country over time that may be associated with the implementation of risk preventive measures and changes in perceptions and concern among the population to be accounted for. The third limitation is related to the overrepresentation of online surveys together with small sample sizes, which can be attributed to the conditions of lockdown and restricted travel that largely constrained the collection of primary data. The present study "partially" addressed this limitation by using a panel dataset collected by means of telephone surveys with 34,219 individuals from five MENA countries, which provided comprehensive coverage and a significantly larger sample size than have been found in existing studies [23,24]. Potential respondents with mild or no interest can ignore or delete e-mailed or online requests to complete an online survey, but a trained telephone interviewer can encourage those who are not enthusiastic about completing the survey to participate [25]. Therefore, it is anticipated that the use of a telephone survey together with the large sample size on which the empirical analysis in this paper is based will mitigate concerns associated with the sample representativeness of webbased and online surveys. The section below sets out the methodology and data used in the empirical analysis. Section 3 reports and discusses the empirical findings, while Section 4 summarises the study and concludes by considering implications for further research. --- Materials and methods --- Theoretical background and empirical model The theoretical underpinnings of this study lean on the health belief model (HBM), which provides a widely accepted framework for understanding and analysing health behaviours [26,27]. The general HBM assumption is twofold: i) health-related behaviours are closely associated with the level of concern and anxiety perceived by an individual as a risk or threat, and ii) it is anticipated that the benefits that an individual would gain from performing such proactive behaviours manage the given threat effectively and outweigh any obstacles to implementing them [28]. Within HBM analytical frameworks, risk prevention or reduction behaviours represent the dependent construct, which is hypothesised to be influenced by individuals' perceptions (e.g. level of concern and perceived severity and susceptibility), and a set of modifying factors comprising demographic characteristics and health-related habits and knowledge. In line with HBM assumptions and considering the nature of the variables included in the panel dataset, concern about COVID-19 infection was used as a proxy for the perceived threat of the pandemic. Compliance with preventive measures by undertaking preventive behaviours, such as washing hands, avoiding mass gatherings and wearing face masks, depended on the level of risk that individuals perceived during different waves of the pandemic. The modifying factors consisted of the following demographic characteristics: gender, age, income, education level and community characteristics captured by the rurality or urbanity of the area in which the individual resides. Thus, the drivers of COVID-19 preventive measures (i.e. the wearing of face masks, social distancing and handwashing) were modelled as a function of infection worries, socio-economic and demographic factors, time invariance factors, and country and administrative-level fixed effects, following Amuakwa-Mensah et al. [29]. The empirical model that explored the relationship between compliance with COVID-19 preventive measures and infection worries is shown in Eq (1): y * ijkt 1<unk>4 b 0 <unk> b 1 Covid worries ijkt <unk> X ijkt <unk> <unk> m jk <unk> Z k <unk> g t <unk> <unk> ijkt<unk>1<unk> where y ijkt is the probability of an individual (i) in community (j) residing in country (k) adhering to COVID-19 preventive measures at time (t). The outcome variables of interest, i.e. the wearing of a face mask, social distancing and handwashing, are binary. It takes the value of 1 if an individual adheres to a specific COVID-19 preventive measure at a point in time, and 0 otherwise. In the case of mask wearing, respondents were asked to indicate whether or not they wear a mask when they are outside the house. For social distancing, respondents were asked to indicate whether they stay at least one metre away from people when they are outside their house. Similarly, respondents were asked to indicate if they wash their hands with soap more often than they did before COVID-19. The term Covid_worries represents the level of concern an individual has about contracting COVID-19. Individuals were asked to indicate how worried they were about being infected with COVID-19. Respondents could choose between the following options: "not worried", "a little worried", "rather worried", "very worried" and "already infected". The term X ijkt is a vector representing socioeconomic and demographic factors, and other control variables such as household size, age, education, marital status, employment status, income quartile and locality (i.e. urban or rural). In addition, survey wave (<unk> t ), country (<unk> k ) and administrative (<unk> jk ) fixed effects were controlled for in the model. The disturbance term is captured as <unk> ijkt. Given the binary nature of the outcome variables, Eq (1) was estimated based on a probit model using the maximum likelihood estimation technique. From Eq (1), while y* ijkt is an unobservable (or latent) outcome, y ijkt was observed, such that: y ijkt 1<unk>4 1 if y * ijkt > 0 0 if y * ijkt 0<unk>2<unk> ( where the zero threshold is a normalisation that is independent of whether the model in Eq (1) includes an intercept or not. Given the latent variable models in Eqs (1) and (2), the model could be reformulated as: p 1<unk>4 Pr<unk>y ijkt 1<unk>4 1<unk> 1<unk>4 Pr<unk>z 0 <unk> <unk> <unk> ijkt > 0<unk> 1<unk>4 Pr<unk> <unk> ijkt <unk> z 0 <unk> 1<unk>4 F<unk>z 0 <unk>3<unk> where is p the probability of an event happening, the function F() is the cumulation density function (cdf) of -<unk> ijkt, and the term z is a vector representing the explanatory variables (i.e. terms X ijkt and Covid_worries), as discussed earlier. The vector <unk> represents the respective coefficient of the explanatory variables in the model. Assuming the error term (<unk> ijkt ) has a standard normal distribution, Eq (3) yields a probit model that is estimated using the maximum likelihood estimation technique. For ease of interpreting the estimated coefficients from this model, the marginal effects after the maximum likelihood estimation were calculated. Formally, the marginal effect of a specific explanatory variable (<unk>) is estimated as follows: @p @z t 1<unk>4 z 0 <unk> <unk> <unk>b t where <unk>() represents the standard normal density function. Based on Eq (1), the coefficient of interest is <unk> 1, which measures the association between compliance with Covid-19 mitigation measures and infection worries. Given the categorical nature of Covid_worries, the option "not worried" was used as the reference category in the estimation. The possible endogeneity of the variable of interest (i.e. Covid_worries) is acknowledged, thus care should be taken when claiming causality in interpreting the results. Potential endogeneity could not be addressed due to data limitation in finding a valid instrumental variable. In addition to the full-sample analysis, sub-sample analysis was also considered focusing on gender, age group, education level, income quartiles, urban-rural and country-specific analysis to account for potential heterogeneity (variation) in relation to the association between adherence to COVID-19 preventive measures and infection worries. --- Data and sampling design This study was based on data from five waves of a Combined COVID-19 MENA Monitor Household survey (CCMMHH) carried out by the Economic Research Forum (ERF) between November 2020 and August 2021 [30]. The CCMMHH survey comprises a suite of panel telephone surveys that are rolled out approximately every two months. Experienced survey research and polling companies in each country conducted the interviews using computerassisted telephone interviewing techniques. Specifically, the data consist of the base wave of the survey (November 2020) and four panel datasets collected in February 2021, April 2021, July 2021 and August 2021. Interviewers sought the consent of respondents on the telephone to their participation in the survey during each survey wave. After introducing the study to the respondents, they were informed that participation was voluntary and asked if they agreed to participate in this survey. The CCMMHH questionnaires covered a range of topics including demographic and household characteristics, labour market and employment characteristics, and income sources and levels. Furthermore, the questionnaires contained a module on attitudes to Covid-19 risks and adherence to the mitigation measures that were implemented during the pandemic in the surveyed MENA countries to contain the spread of the virus, which comprised wearing a face mask, maintaining social distancing and washing hands. Specific questions on all the modules of the questionnaire are provided by the Economic Research Forum, which is responsible for the Combined COVID-19 MENA Monitor Household survey (CCMMHH) [30]. The MENA countries covered by the survey are Egypt, Jordan, Morocco, Tunisia and Sudan. These countries did not appear in all the waves; thus, each country has a different baseline and follow-up periods. For each country and wave, the corresponding sample is shown in S1 Table. A stratified sampling design was used based on mobile operators' country-specific market shares. The sample was designed to cover at least 2,000 unique households and individuals, focusing on mobile phone users aged 18-64 and using random digit dialling within the range of valid numbers. The selected number is dialled up to three times and unanswered phone numbers or those who answered but could not complete the survey were dropped from the study. More information on the sampling design and response and attrition rates is provided by the Economic Research Forum, which is responsible for the CCMMHH survey [30]. Respondents from previous waves were contacted again for a follow-up interview only if they had agreed to this in the earlier wave. Similarly, up to three call attempts were made to respondents for the follow-up interview and, in some cases, second and family/friends' numbers were also contacted in an attempt to reach the respondent. In the event that an individual refused to respond or was unreachable for the follow-up interview, the individual was replaced by a new person who was randomly selected in accordance with the base wave protocols. In order to reduce sampling bias, an inverse probability weighting was applied based on the following: a) telephone operators and their market shares, b) number of telephones by operator for individuals and household members, and c) representative in-person survey data with comparable demographic and household characteristics. More details can be found on the ERF website in this link: http://www.erfdataportal.com/index.php/catalog/230/download/ 3459. The descriptive statistics of the variables used in this study are shown in Table 1. Generally, a greater proportion (around 80%) of individuals in the MENA region adhered to COVID-19 containment measures such as social distancing, mask wearing and handwashing. However, Sudan had a relatively low percentage of individuals adhering to social distancing (56%) and wearing a mask (62%). The average age of the sample was around 36 years, and the household --- Empirical results --- Full-sample estimates Table 2 presents the marginal effect based on the three probit model estimations, which were estimated to examine the association between the level of individuals' concern about contracting COVID-19 and their compliance with their government's risk prevention and mitigation measures. A deeper look at the estimates related to the independent variable (concern about infection) in the estimated models for maintaining social distancing, using a face mask and frequent handwashing, revealed two interesting findings. First, the results demonstrated that the adoption of COVID-19 mitigation measures was positively associated with the levels of perceived concern about infection, while the coefficients of this variable in three estimated models were positive and highly statistically significant at the 1% level. Second, the marginal effect of adherence to COVID-19 mitigation measures in the three models was found to rise as the level of individuals' concern grew from "a little worried" to "very worried", and then markedly decreased when individuals were infected with the virus. For instance, in the social distancing model, the value of the concern coefficient rose from 0.09 for the variable category "a little worried" to 0.11 and 0.12 for the variable categories of "very worried" and "rather worried" respectively, and then nearly halved (0.06) for the "already infected" category. Similarly, the values of the coefficient rose from 0.08 and 0.09 for the "a little worried" category of the sample in the face-mask wearing and handwashing models to 0.09 and 0.11 respectively for the "rather worried" category. After reaching their maximum value for the "very worried" category (0.10 and 0.13), they then fell below the baseline level of the reference group (not worried) for the "already infected" category of the sample. In addition, the parameters of the individual attributes (age, gender and education) and household size reported in Table 2 indicated that demographic and household characteristics were important determinants of compliance with public health measures. In particular, females were found to be significantly more likely to comply with the three mitigation measures than their male counterparts. Likewise, the results revealed that adoption of protective behaviours increased with age. In contrast, the association between formal educational attainment and compliance with COVID-19 mitigation measures was rather tenuous and to some extent paradoxical. For instance, individuals with a basic level of education were found to be significantly less likely to use face masks, practise repeated handwashing, and maintain social distancing during the COVID-19 pandemic than others with a less than basic level of education. Nevertheless, the results highlighted that individuals with higher educational degrees were significantly more likely than those with a less than basic level of education to wear face masks and practise repeated handwashing during the COVID-19 pandemic. Despite the statistical insignificance of the coefficients related to social distancing and handwashing, in terms of household factors, the negative sign of the coefficient of household size in the three models suggested that a larger household size generally hampered the ability of individuals to comply with mitigation measures. Finally, no statistically significant evidence was found of the impact of the characteristics of the community (rural vs. urban areas) in which an individual resides on compliance with COVID-19 preventive and mitigation measures. --- Heterogeneity analysis Given the possible variation in the association between concern about COVID-19 infection and compliance with mitigation measures, this relationship was explored across the respondents' gender, age group, income quantiles and levels of educational attainment. A countrylevel analysis was also considered in this section, given the potential variation in the association between concern about COVID-19 infection and compliance with safety practices, due to differences in the policies implemented in the various countries. ) show the marginal effects of concern about infection on the probabilities of adhering to COVID-19 mitigation measures by gender. Both the sign and magnitude of coefficients related to the association between the level of an individual's concern and adherence to repeated handwashing and maintaining of social distancing were qualitatively similar to those reported by the pooled model (Table 2). However, the results notably showed that the marginal effect of concern about infection on the use of face masks was higher for men than women across all categories of the independent variable, implying that men were more likely to adhere to these particular proactive behaviours. --- Age group. Fig 2 (and S3 Table) present the marginal effects of the probit model by age group. Consistent with the full-sample estimates, the results confirmed that individuals who felt a high level of concern about infection (very worried) across all age groups were more likely to comply with Covid-19 mitigation measures, compared with those who perceived little or moderate levels of concern. A comparison of the magnitudes of the coefficients related to the three mitigation measures across age groups revealed that individuals belonging to the age group 35 years and below were more likely to follow all the mitigation measures, compared with other individuals in the age groups over 35 years. Interestingly, the results showed that younger individuals (under 35) were more compliant with recommendations for social distancing and using face masks than the elderly (over 60). --- Educational attainment. Fig 3 (and S4A and S4B Table) display the marginal effects of concern about COVID-19 infection on the probabilities of compliance with COVID- 19 mitigation measures across levels of education. The results revealed that the association between concern about infection and adherence to COVID-19 mitigation measures was highly significant in that the marginal effect in relation to the three preventive measures rose, with a few exceptions, as the level of education increased from less than basic and basic education to secondary and higher education levels. ) report the estimated marginal effect of concern about COVID-19 infection on individuals' compliance with mitigation measures across income groups. The results showed that individuals in the third and fourthincome quantiles were more likely to comply with the three mitigation measures than those in the lower income quantiles, providing a strong case for the effect of household income on individuals' compliance with COVID-19 mitigation measures. In particular, the marginal effects of compliance with the mitigation measures in relation to the use of face masks were highest in absolute terms, indicating that the higher a household income was, the greater the probability of an individual following government recommendations regarding the wearing of face masks in public places during the pandemic. 3.2.5. Country-level analysis. Tables 3 and4 summarise the estimated marginal effect of concern about COVID-19 infection on individuals' compliance with mitigation measures by country. The results showed distinct differences between the surveyed countries in relation to the influence of concern about infection on public compliance with the implemented preventive measures, with the strongest association being observed in Tunisia and Sudan and the lowest in Jordan and Morocco. A noteworthy observation from Tables 3 and4 is that except for Morocco, the results confirmed the findings presented in Table 2 regarding the "first-up-thendown" trend of the odds of compliance with COVID-19 mitigation measures, which increased with individuals' level of concern about infection from "a little worried" to "very worried", and then fell when individuals were infected by the virus. --- Discussion Overall, the empirical results provided plausible evidence of the impact of risk perception on individuals' risk-related behaviour. The probit model estimates revealed a positive and statistically significant association between level of concern about contracting COVID-19 and individual compliance with the recommended measures implemented by the governments of the surveyed countries for risk prevention and mitigation. This is consistent with the findings of previous studies in this field, which suggest that perceived risks and fears of infection are dominant predictors of compliance with public health measures [e.g. 24,31]. In other words, individuals with higher levels of fear and risk perception are often more concerned, anxious and stressed, and therefore more likely to follow and comply with preventive measures such as handwashing, using face masks and social distancing. Interestingly, adherence to COVID-19 mitigation measures were found to follow a "firstup-then-down" trend, while compliance with the three mitigation measures went hand in hand with the level of individuals' concern about contracting COVID-19, but then markedly decreased after individuals contracted the virus. This finding supports those of Lio et al. [32] and Binte et al. [33], who revealed that individuals with a positive COVID-19 history were less likely to adhere to mitigation measures compared with uninfected individuals who have no prior history of infection. This tends to imply that individuals who were infected and then recovered perceived the disease as posing less risk, which subsequently reduced their protective behaviour [34]. Another explanation is that individuals who were infected may have had lower levels of compliance because of the sense of security that recovery from infection may provide [35]. However, research on previous disease outbreaks indicates that sharp decreases in risk perception can contribute to reinfections and the emergence of much severer subsequent waves of infection that might cause high mortality rates and have serious economic effects [36]. For instance, previous research has shown that lower levels of compliance with risk mitigation measures together with relaxed or no lockdown measures contributed to an increase in the rates of reinfection in several countries. These were estimated by Ribeiro Xavier et al. [37] to be as high as 40% in South Africa, and by Hoang [38] to be 14% in China and 31% in Korea. --- Table 3. Marginal effect of individuals' concern about COVID-19 infection on compliance with mitigation measures by country. --- Concern about infection | --- Jordan In line with the recent literature on COVID-19 risk perception and adoption of preventive measures [24,39], the results of the present study showed that individual attributes (age, gender and education) and household size were key predictors of individuals' compliance with public health measures. The finding that females were significantly more likely than their male counterparts to adhere to the three outcome measures is consistent with that of Ning et al. [40], who point out that women generally tend to have a higher level of concern and fear of infection, and a greater sense of responsibility and willingness to protect society, making them keener and more likely to engage in recommended health behaviours than men. Likewise, the results revealed that older adults were more likely than younger adults to adopt protective behaviours. This supports the findings of Bronfman et al. [41], who illustrate that older adults perceive themselves to be more vulnerable to disease infections, especially in the presence of comorbidities, which motivates them to embrace protective behaviours, compared with younger adults, to avoid worse outcomes of COVID-19 infections. In relation to the role of educational attainment, individuals with a basic level of education were found to be significantly less likely to comply with mitigation measures than others with a less than basic level of education. In this respect, Lieberoth et al. [42] report similar findings, while their results revealed that people with lower education levels tended to feel more anxious and stressed by the COVID-19 pandemic, and thus were more likely to follow the recommended preventive measures. However, the present study found evidence that individuals with higher educational degrees were significantly more likely than those with less than basic levels of education to adhere to the recommended mitigation measures during the COVID-19 pandemic. This finding is in agreement with Oyetunji et al. [43], who point out that having more than twelve years of formal schooling is positively associated with practising more than one protective measure to prevent infection. In general, the results related to household size in the three models indicated that a larger household size hampers individuals' ability to comply with mitigation measures. In this respect, Ye et al. [44] investigated individual and household factors influencing the adoption of preventive behaviours among a large sample of households from eight provinces in China, and found that members of larger households had limited capacities to adopt COVID-19 preventive measures. In particular, the statistically highly significant coefficient of this variable in the second model (using a face mask), in contrast to the handwashing and social distancing models, can be attributed to the nature of this behaviour in the sense that several MENA countries experienced shortages of medical masks during the early stages of the pandemic [45]. It is also the only preventive behaviour among the three investigated behaviours that involves a direct financial cost [46], which may be unaffordable for larger households. Finally, no statistically significant evidence was found of the impact of the characteristics of the community (rural vs. urban areas) in which an individual resides on compliance with COVID-19 preventive and mitigation measures. Turning to the results of the heterogeneity analysis, both the sign and magnitude of coefficients in the estimated gender-based models were qualitatively similar to those reported by the pooled model, and conformed with previous studies showing that women are more likely than men to follow proactive health measures during public health crises [41]. One possible explanation for this is that males' usual social and professional or labour activities lead them to leave home more often and socialise more frequently, which had implications for their adoption of preventive behaviours [40]. However, the results showed that the association between concern about infection and use of face masks was stronger among men than women across all categories of the independent variable. A study by Balkhi et al. [47] in the context of Pakistan ascribed similar findings to certain cultural and environmental factors as men are most often the breadwinners in the household, and thus had to comply with mandatory measures relating to maintaining social distancing and wearing face masks in workplaces and public places. In the same vein, Alghalyini et al. [48] demonstrate that women in Saudi Arabia, a country in the MENA region, were less likely than men to leave their home during the pandemic and conducted more work-related and household tasks from home (e.g. working from home and online shopping). This minimised their interactions with others in public places and reduced the likelihood that they would need to comply with mandatory measures of using face masks in public spaces and workplaces. The age-differentiated estimates confirmed the full-sample results that higher levels of concern about infection were associated with greater likelihoods of complying with Covid-19 mitigation measures across all age groups. Notably, the finding that younger individuals (under 35) were more likely to follow all the mitigation measures than individuals over 35 slightly contradicts previous studies, which show that middle-aged individuals reported higher levels of compliance with COVID-19 preventive behaviours than their younger counterparts [18]. However, the results showed that the elderly (over 60) were less compliant with social distancing and face mask recommendations than younger individuals (under 35). This finding is in agreement with the findings of Alahdal et al. [49], showing a significant difference in the level of compliance with COVID-19 proactive measures with better practices observed among younger individuals (18-49 years) compared with those aged over 60. This was attributed to their greater use of various social media channels compared with the elderly. In terms of the influence of educational attainment on the association between concern about infection and adherence to COVID-19 mitigation measures, it was found that, with a few exceptions, the likelihood of an individual's compliance with the three preventive measures increased with the level of the individual's educational attainment. This finding is in agreement with previous studies indicating that lower education levels are barriers to compliance with COVID-19 preventive measures [50], and that individuals with higher education levels are more aware of and more anxious about the COVID-19 pandemic, which increased the probability of their adoption of COVID-19 preventive measures and precautionary practices [51]. A comparison across income groups revealed that individuals belonging to higher income quantiles (i.e. the third-and fourth-income quantiles) were more likely to comply with the three mitigation measures than those in the lower income quantiles. In particular, this was the case for the use of face masks, where income level was a highly significant predictor of individuals' adherence to government recommendations regarding the wearing of face masks in public places during the pandemic. This finding is in agreement with the wide array of studies in this field that have found evidence of the role of income in predicting individuals' adherence to COVID-19 proactive and mitigation measures [18,52]. For instance, El-Shal and Moustafa [53] show that economic hardship moved people in MENA away from adhering to COVID-19 mitigation measures as they became less concerned about being infected with COVID-19, perceiving that they had little to lose given their already precarious existence. Alkhaldi et al. [54] point out that low-income households often encounter significant barriers in relation to the adoption of preventive measures because of their crowded living situation, involvement in the informal labour market, and inability to afford the cost of hygiene products, which collectively makes them less able and willing to comply with these mitigation measures. Shahin and Hussein [52] illustrate that the purchasing of mitigation and precautionary equipment during the COVID-19 pandemic (e.g. alcohol, detergents and face masks) represented an additional monetary burden for many Egyptian households, whose income and financial status had already worsened due to the adverse impacts of the pandemic on employment, incomes and labour markets. Finally, the results revealed distinct differences across the surveyed countries regarding the association between the level of concern about contracting COVID-19 infection and public compliance with the preventive and mitigation measures. While many previous studies on MENA countries have often regarded the region as a single homogeneous unit [e.g. 14,15], this finding underscores the issues that may result from ignoring the fact that MENA is a highly diverse region, incorporating low, middle and high-income countries at different stages of socio-demographic transition. It also contains socio-politically stable countries and others with ongoing conflicts and emergencies. Neglecting these differences precludes the possibility of analysing the relationship between concern and compliance from a deeper perspective and makes a generalisation of the results questionable. In this regard, existing evidence attributes heterogeneities in compliance with COVID-19 mitigation measures between countries to a suite of governance and institutional factors (e.g. the stringency of the measures, their length of implementation and the degree of social trust in governments), and a range of socioeconomic factors (e.g. the demographic composition of the population, including age and gender, social capital, the degree of ethnic diversity, income level), and political beliefs, and other cultural beliefs and social characteristics among the population [55,56]. Therefore, heterogeneities among MENA countries in relation to the varying levels of the infrastructures of health systems have arguably created specific challenges for individual countries and influenced the provisions of public health policies during the pandemic, led to uneven responses by governments in terms of the implementation of risk management measures, and largely determined people's compliance with the enforced containment measures [57]. Despite the consistency and relevance of the present study's results, three limitations of the present study should be noted and considered in future research. First, the major limitation is that the CCMMHH survey was designed for other purposes and without reference to the empirical model, which indeed limited the efficacy of this model. For instance, psychological models of behaviour change, including the HBM that formed the theoretical underpinnings of this study, suggest that individual adoption of protective behaviours rests on a much wider array of factors than just risk perception. Owing to the limitations of the survey data, some important predictors of behaviour were excluded. Future research would benefit from extending these results and fleshing out the model by accounting for psychological predictors (e.g. risk attitudes, beliefs, self-efficacy and perceived behavioural control) and contextual variables (e.g. economic, socio-political and temporal factors) that may influence public ability and willingness to comply with risk preventive mitigation measures. A second limitation of this study is that the sample was obtained through a telephone panel survey, which subjects the results to the biases that this context might entail. Finally, the third limitation is related to potential endogeneity of the concern about infection variable, which prevents a claim of a clausal relationship. --- Conclusions and policy implications Using

This study investigates the association between individuals' concern about contracting COVID-19 and their compliance with recommended preventive and mitigation measures, namely wearing face masks, maintaining social distancing and handwashing, in the context of the Middle East and North Africa (MENA) region. The empirical analysis is based on a panel dataset from the Combined COVID-19 MENA Monitor Household Survey, which was carried out in Jordan, Morocco, Sudan, Tunisia and Egypt. Applying a probit estimation technique, a positive and statistically significant association was found between the level of COVID-19 worries and individuals' compliance with the mitigation measures. Notably, the results revealed that this association followed a "first-up-then-down" trend, showing that compliance with the three mitigation measures rose as individuals' worries about contracting the virus increased, and then markedly decreased after they had been infected. Sociodemographic characteristics contributing to lower levels of compliance included being male, being over 60, having lower levels of education and having a lower household income. A cross-country analysis revealed remarkable differences between the five countries, with the strongest association between COVID-19 concerns and adherence to mitigation measures observed in Tunisia and Sudan, and the weakest association seen in Jordan and Morocco. Policy implications are outlined for effective risk communication and management during disease outbreaks and public health emergencies to encourage appropriate public health behaviours.

CCMMHH survey was designed for other purposes and without reference to the empirical model, which indeed limited the efficacy of this model. For instance, psychological models of behaviour change, including the HBM that formed the theoretical underpinnings of this study, suggest that individual adoption of protective behaviours rests on a much wider array of factors than just risk perception. Owing to the limitations of the survey data, some important predictors of behaviour were excluded. Future research would benefit from extending these results and fleshing out the model by accounting for psychological predictors (e.g. risk attitudes, beliefs, self-efficacy and perceived behavioural control) and contextual variables (e.g. economic, socio-political and temporal factors) that may influence public ability and willingness to comply with risk preventive mitigation measures. A second limitation of this study is that the sample was obtained through a telephone panel survey, which subjects the results to the biases that this context might entail. Finally, the third limitation is related to potential endogeneity of the concern about infection variable, which prevents a claim of a clausal relationship. --- Conclusions and policy implications Using a panel dataset based on five waves of the Combined COVID-19 MENA Monitor Household Survey (CCMMHH) conducted in Jordan, Morocco, Sudan, Tunisia and Egypt between November 2020 and August 2021, this study examined the association between individuals' worries about contracting COVID-19 and their adherence to the recommended preventive measures of wearing a face mask, maintaining social distancing and handwashing. The empirical findings provide evidence about community practices around the use of risk prevention and mitigation measures during public health crises in settings with different infection incidences and response policies, which should inform policies that aim to mitigate the transmission of infectious diseases and combat their negative externalities. Specifically, a number of conclusions and policy implications can be drawn for effective risk communication and management during disease outbreaks and public health emergencies to encourage appropriate public health behaviours. First, the overall results provide plausible evidence of the impact of risk perception on individuals' risk-related behaviours, with the probit model estimates revealing a positive and statistically significant association between level of concern about contracting COVID-19 and individual compliance with the recommended measures implemented by the governments of the surveyed countries for risk prevention and mitigation. Government interventions aimed at promoting the adoption of risk mitigation measures during pandemics and disease outbreaks should tailor messages more effectively to communicate health risks to susceptible segments of the population and enhance their perceptions of threat and self-efficacy, and thereby encourage them to perform certain behaviours. In the same vein, government campaigns should also focus on boosting public attitudes and knowledge by nudging them toward trusted sources of information about causes of infection and associated health risks to prevent negative outcomes of conspiracy theorising in times of public health crises. To this end, building trust in government authorities is essential in order to influence the perception of risk and promote the adoption of preventive public health measures. Second, the results highlight a decreasing trend in the association between concern about COVID-19 infection and compliance with preventive and mitigation measures after an individual has contracted the virus. This implies that worries about infection decreased as the pandemic continued, and that when people became more familiar with the pandemic, including by contracting the virus, they perceived fewer health risks and thus adopted less strict hygienic behaviours. This should be a cause of concern with regard to the effectiveness of social protection measures that the governments of the surveyed countries implemented to mitigate the economic impacts of their mitigation measures, which seem to have created a dilemma between complying with the preventive measures and maintaining a source of income. In connection with this, a look at public policy discussions in relation to the COVID-19 pandemic in MENA countries, especially at the beginning of the outbreak, shows that they were dominated by health concerns, whereas little focus was given to the economic side-effects of the implemented mitigation measures, which might be riskier than the virus. Such false trade-offs, which prioritised health and ignored the economy, seem to have reduced acceptance of the implemented measures among people with lower incomes and other marginalised groups in the population. Therefore, these findings imply that risk mitigation interventions and communication approaches need to shift from a context of emergency and uncertainty to one that encourages sustainable and habitual behaviours. Considerations within government interventions must address the wider impacts of disease control measures by including dimensions beyond biomedical risks that account for the social and economic effects of these measures on the population over time, as these effects may act as a deterrent to public compliance with risk prevention and mitigation measures. Third, the results clearly point to the existence of remarkable variations between the individuals in the surveyed sample in relation to the determinants of their compliance with COVID-19 mitigation measures based on their age, gender and other socio-demographic and economic characteristics. A deeper understanding of the perceptions and attitudes towards mitigation behaviours and their underlying drivers among different socio-demographic groups is a key element in bringing about behavioural changes and promoting compliance with these measures. In conjunction with this, more research is needed to understand the communication preferences and trusted sources and channels of information in order to design tailored interventions for each target group. In this respect, it is essential to map the key trusted influencers for each group and engage them in crafting tailored communication and nudging messages to ensure that desired behavioural changes are effectively accomplished. Such strategies are more likely to be successful if they co-are designed through community-centred approaches that acknowledge local realities and meaningfully involve the target groups to identify and implement locally appropriate mitigation measures. Fourth, the cross-country analysis reveals regional heterogeneities among the five surveyed countries in relation to the influence of COVID-19 concerns on public adherence to the implemented mitigation measures. This offers an opportunity for regional cooperation and collaboration between policymakers in public health sectors within the sampled countries. Specifically, policymakers in these countries can capitalise on these heterogeneities to develop cross-government approaches and create platforms for ongoing dialogue among stakeholders to promote a sharing of experience about effective approaches and best practices in order to reinforce public acceptance and people's willingness to adhere to preventive and mitigation measures. --- Supporting information S1 Table. Number of respondents to the CCMMHH survey from each country across the survey waves. Source: CCMMHH survey [44]. (PDF) --- S2 Table. Marginal effect of individuals' worriedness about COVID-19 infection on compliance with mitigation measures by gender. | Reference group is "not worried". Standard errors in parentheses *** p<unk>0.01, ** p<unk>0.05, * p<unk>0.1. We controlled for household size, urban, gender, education, marital status, employment status, income quartile, wave, country and administrative fixed effect in all the models. (PDF) --- S3 Table. Marginal effect of individuals' worriedness about Covid-19 infection on compliance with mitigation measures by age. | Reference group is "not worried". Standard errors in parentheses *** p<unk>0.01, ** p<unk>0.05, * p<unk>0.1. We controlled for household size, urban, gender, education, marital status, employment status, income quartile, wave, country and administrative fixed effect in all the models. (PDF) S4 Table. (a and b). Marginal effect of individuals' worriedness about COVID-19 infection on compliance with mitigation measures by education level. | Reference group is "not worried". Standard errors in parentheses *** p<unk>0.01, ** p<unk>0.05, * p<unk>0.1. We controlled for household size, urban, gender, education, marital status, employment status, income quartile, wave, country and administrative fixed effect in all the models. (PDF) S5 Table. (a and b). Marginal effect of individuals' worriedness about Covid-19 infection on compliance with mitigation measures by income quantile. | Reference group is "not worried". Standard errors in parentheses *** p<unk>0.01, ** p<unk>0.05, * p<unk>0.1. We controlled for household size, urban, gender, education, marital status, employment status, income quartile, wave, country and administrative fixed effect in all the models. (PDF) --- Author Contributions Conceptualization: Assem Abu Hatab, Franklin Amuakwa-Mensah. --- Data curation: Assem Abu Hatab. Formal analysis: Assem Abu Hatab, Franklin Amuakwa-Mensah. Investigation: Assem Abu Hatab, Lena Krautscheid, Franklin Amuakwa-Mensah. Methodology: Assem Abu Hatab, Franklin Amuakwa-Mensah.

This study investigates the association between individuals' concern about contracting COVID-19 and their compliance with recommended preventive and mitigation measures, namely wearing face masks, maintaining social distancing and handwashing, in the context of the Middle East and North Africa (MENA) region. The empirical analysis is based on a panel dataset from the Combined COVID-19 MENA Monitor Household Survey, which was carried out in Jordan, Morocco, Sudan, Tunisia and Egypt. Applying a probit estimation technique, a positive and statistically significant association was found between the level of COVID-19 worries and individuals' compliance with the mitigation measures. Notably, the results revealed that this association followed a "first-up-then-down" trend, showing that compliance with the three mitigation measures rose as individuals' worries about contracting the virus increased, and then markedly decreased after they had been infected. Sociodemographic characteristics contributing to lower levels of compliance included being male, being over 60, having lower levels of education and having a lower household income. A cross-country analysis revealed remarkable differences between the five countries, with the strongest association between COVID-19 concerns and adherence to mitigation measures observed in Tunisia and Sudan, and the weakest association seen in Jordan and Morocco. Policy implications are outlined for effective risk communication and management during disease outbreaks and public health emergencies to encourage appropriate public health behaviours.

Introduction Rohingyas are a group of people from the Rakhine state of Myanmar who fled their country due to persecution and expulsion [1], to countries such as Bangladesh, Thailand and Malaysia. In fleeing their homeland, they lost their loved ones, their homes, and their possessions. They have a higher prevalence of communicable diseases including vaccine-preventable and water-borne diseases such as cholera, bloody diarrhoea, typhoid and hepatitis E [2,3] due to poor hygiene conditions in the camps and during their flight from conflict [4]. The population also suffers from a high prevalence of non-communicable diseases (NCD) such as hypertension and diabetes [3] which often goes undiagnosed or undermanaged. Additionally, the mental health of the refugees and asylum seekers is a concern because of the abuses and other forms of exploitations [3] and traumas they experienced. Studies have found mental health issues such as post-traumatic stress disorder (PTSD), depression and suicidal thoughts to be prevalent among Rohingyas [3,5]. Their difficulties do not end after fleeing their country. A new set of obstacles awaits them when they arrive in a new country. When they reach the shore or the boundary of a new country, they often realise that they may not be able to live the better life they had expected. Basic needs such as safe access to food, water, sanitation, shelter and healthcare services, may not be readily available to this population [6]. The physically and mentally agonising experiences associated with their forced migration and restricted or non-access to healthcare services in Malaysia contributes to the deterioration of their health [7] and potentially impacts the population of the host country. Rohingyas live in Malaysia as refugees and asylum seekers. Refugees are people who have fled their own country because of a legitimate fear of persecution for various reasons, are outside their country, and are unable or unwilling to avail themselves of the protection of that country [8]; while asylum seekers are those whose claims for protection have not yet been decided by the country to which they have submitted it [9]. Refugees referred to in this article are those who possess a UNHCR registration card; while asylum seekers are those who are not documented or have their registration accepted by the UNHCR Malaysia, and do not possess the UNHCR registration card. Although as refugees and asylum seekers, both Rohingya women and men have faced violence during the conflict in Myanmar, in-transit, and in Malaysia, their experiences of violence and its impact are different [10]. Rohingya women face difficulties with safety and security, sexual and reproductive health (SRH) issues such as safer pregnancy and delivery, marriage practices, gender-based violence (GBV) and other issues [11]. Most studies conducted to date on refugee women in Malaysia have explored issues from the public health perspective, which includes psychology and psychiatry [12][13][14][15] but not social work, despite its importance. Therefore, it is important to explore the challenges that Rohingya women refugees and asylum seekers face with respect to Malaysian healthcare services, and the role that social workers, within the context of healthcare and with other professionals, can potentially play in providing their services. The investigation will inform the development of policies and practices which align with the anti-oppressive principles of reducing their barriers to healthcare. This article reviews past studies to explore the barriers Rohingya women in Malaysia face in accessing healthcare services, and some of the reasons why Malaysian social workers cannot extend their services to the refugee and asylum-seeking population. It then briefly introduces the anti-oppressive social work theory as a backdrop to the arguments for social workers in Malaysia to provide their services to refugees and asylum seekers, which is put forth in this article. The research design, and methods used to conduct the research, are subsequently presented. To understand the physical and financial barriers to healthcare that Rohingya women face, and the services medical social workers provide in the country, the article presents primary data collected from refugees and asylum seekers, in addition to social workers and others who provided their services to these people. Building on these findings, the article offers a discussion on the physical, mental and financial challenges Rohingya women face, and medical social work services that are provided to refugees and asylum seekers in the country. Additionally, it offers recommendations and concludes with an argument for medical social work practice with this population in Malaysia. --- Barriers to Healthcare That Rohingya Women Refugees and Asylum Seekers Face in Malaysia In Malaysia, refugees and asylum seekers are not legally permitted to work and do not have access to free public healthcare or education. The country is not party to the 1951 Convention Relating to the Status of Refugees and its 1967 Protocol [16]. Therefore, a legal framework to distinguish a refugee or an asylum seeker from another undocumented migrant does not exist [17]. This leaves refugees, asylum seekers and stateless persons viewed as people without documents who can be fined, arrested, detained, imprisoned, punished and deported [17]. The challenges they encounter in Malaysia will lead them and their children to remain trapped in the cycle of poverty and other forms of disadvantages. The Rohingyas comprise the largest group of people among Malaysia's registered refugee groups, followed by the Chins from Myanmar [18]. They are also one of the world's most persecuted ethnic groups [19], whose people are highly vulnerable to various forms of deprivation. They are stateless because their citizenship in Myanmar has not been recognised, depriving them of social, political, educational, and healthcare rights [20]. To make matters worse, the lack of legal protection, the normalisation of violence within refugee groups, and the difficulty in accessing protection and justice for refugee women in Malaysia, contributes to an environment where they are particularly vulnerable to sexual and gender-based violence (SGBV) [21]. These challenges potentially lead them to life-threatening healthcare issues such as sexually transmitted diseases (STDs), unwanted pregnancies, menstrual disorders and psychological trauma [22]. These diseases often go undiagnosed due to many factors which include cultural barriers [23,24], financial constraints, and lack of access to healthcare services [24]. Although there are several non-governmental organizations (NGO)s and/or volunteerrun clinics for refugees in Malaysia [25], the high charges in hospitals and private clinics are the reason why refugees frequently visit refugee clinics. However, refugee clinics can only provide basic healthcare services; when refugees need advanced healthcare services, they will still need to be referred to hospitals. Furthermore, undocumented women who seek healthcare assistance in hospitals face possible arrest in Malaysia. Health Equity Initiatives (HEI) reported incidences whereby asylum-seeking pregnant women who were admitted to a public hospital were told that upon delivery of their babies, they would be sent to a detention centre, which caused great anxiety to the women and their husbands on top of their usual fear associated with childbirth [13]. Fear of arrest, detention and deportation deterred these women from seeking professional healthcare assistance to deliver their babies, and may also lead to unsafe abortion or post-natal infection that risks maternal mortality [13]. --- Why Social Workers Could Not Provide Their Services to These Refugees and Asylum Seekers NGOs and individual volunteers assist and support refugees in Malaysia to a certain extent. In the United States, for example, social workers engage and work with refugees from the time of the refugee's arrival in the country. Each state in the country has its own social workers licensed to practice [26], and a board to regulate state licensing [27]. They provide basic needs for at least 30 days and perform case management and referrals to various social service providers. In New Zealand, the manager of the Refugee Resettlement Centre conducts an orientation programme, which includes language classes, health screening, and mental health support before resettling refugees into the country [28,29]. Similarly, in Western Australia, refugees go through post-arrival health screening, which includes trauma-informed and culturally appropriate healthcare services by a multidisciplinary team that includes social workers [30]. However, it must be noted that asylum seekers are predominantly smuggled into Malaysia [31], unlike those in the United States, New Zealand or Australia where UNHCR and other agencies located around the world assist in resettling refugees into the three countries from various other countries [32,33]. This enables the allocation and provision of systematic and adequate resources that includes social work services for asylum seekers or refugees. In Malaysia, social workers are under the purview of the Department of Social Welfare within the Ministry of Women, Family and Community Development (MWFCD), while medical social workers are under the Ministry of Health (MoH). Medical social workers, of which the vast majority work in hospitals, perform administrative duties in addition to their duties to serve patients in need [34]. The roles these social workers play in the country are somewhat unclear because, inherently, social work is perceived to be almsgiving work, which does not require specialised training or qualification [35]. Other health professionals have been reported as performing some of the tasks of these social workers [35]. Due to the heavy workload, stressful nature of their jobs, inadequate workforce, and more importantly, a mandate, social workers are not able to provide even basic services such as preliminary assessment and referrals for asylum seekers and refugees who seek medical attention in public hospitals. --- Anti-Oppressive Social Work The anti-oppressive concept works on the premise of oppression. Oppression is defined as involving "relations of domination that divide people into dominant or superior groups and subordinate or inferior ones" [36]. The "superior" excludes the "inferior" from the available social resources through a systematic devaluation of the attributes and contributions of the latter [36]. Anti-oppressive social work was one of the approaches that fell under the umbrella of critical social work [37]. The social work profession, which is defined as one that "promotes social change, problem-solving in human relationships and the empowerment and liberation of people to enhance well-being", strongly resonates the idea of anti-oppressive practice (AOP) [38]. Scholars have used and contributed to the anti-oppressive approach in both research and practice in the field of social work. At the micro-level category of practice, the limitations of anti-oppressive perspective in social work has been discussed and the use of critical consciousness was emphasized to fill its gaps [39]. On the other hand, an international comparative framework was used at the macro-level category of research to contextualise the debates and controversies surrounding AOP theorization in social work [38]. The anti-oppressive approach has also been utilised in various studies on refugees, including on the role social workers can play through their approach during the refugees' initial settlement period in a host country. Two studies [40,41] explored an anti-oppressive practice model to support queer and trans refugees in Canada using critical intersectionality analysis. Another study [37] reflected on their practice of working with asylum seekers in Australia by delving into the experience of oppression and privilege in addition to a commitment to personal, cultural, and structural change. Negative stereotypes among others is a manifestation of oppression that constrains the relations between the refugees and their host society as a whole [40]. Knowing the circumstances of the immigrants, and equipped with the information needed to counter stereotypes, social workers can engage civil societies to challenge the misrepresentation or distortion of facts inflicted upon this group of marginalised people [40]. --- Research Design The research design employed to conduct this research was qualitative case study. Case studies are reported to have great value in building emergent knowledge in social work [42]. Primary data were obtained from the informants, who included Rohingya women refugees and asylum seekers, medical social workers, medical officers, volunteer workers/activists, refugee organization officers, and a mental health service provider (see Table 1). The Rohingya Women Development Network (RWDN), a Rohingya women NGO in Malaysia, collected data from Rohingya women informants. The organisation works with and/or for Rohingya women and girls in a wide variety of programmes which include advocacy, education, and women's empowerment. This research used focus group discussions with two groups totalling 10 Rohingya women and in-depth interviews with another 12 Rohingya women. Key-informant interviews were conducted with four medical social workers and two medical officers in a public hospital in Malaysia. Other keyinformants included two volunteer workers/activists, two refugee organisation officers, and one mental health service provider for refugees. The age of Rohingya women informants ranged from 21 to 50 years. All, except for two, had at least one child. They were all born in Myanmar and had fled the country to Malaysia, except for one. Most (n = 18) fled Myanmar and arrived in Malaysia using Thailand as a transit country, while others (n = 4) used Bangladesh and Thailand. The FGDs took approximately one and a half hours to two hours, while the KIIs ranged from half an hour to one and a half hours per session. The first author translated and transcribed interviews with Rohingyas together with Rohingya translators/transcribers who were proficient in both Rohingya and English languages. The authors recruited the translators/transcribers not only based on their language proficiency but also on their knowledge and years of experience of working with Rohingyas for several international and local organisations and the government. The translated and transcribed interview and focus group discussion transcripts were randomly selected to be re-assessed to ensure that the translation accurately represented the informant's statements and their meaning. This research used a thematic analysis method to organise and analyse the data collected. Thematic analysis is a method within qualitative data analysis that is conducted through identifying, analysing, and reporting patterns (themes) within data [43]. The researchers adapted and used the phases of thematic analysis [44] to perform data analysis for this research: (1) familiarize yourself with your data, (2) generate initial codes, (3) look for themes, (4) review themes, (5) define and name themes, and (6) write the thesis. --- Ethics All informants of this research were given an information sheet about the research. The informants then signed a consent form to participate in the research and allow the researchers to publish the information they provided without any of their identifying details (unless otherwise agreed). The informants were informed of their rights to refuse any questions or withdraw from the research at any point. Furthermore, the informants were advised that their participation in the research may not give them any direct benefits. A safe and comfortable space was provided to Rohingya women informants for the research. The researchers also employed a female Rohingya interviewer/facilitator and a note-taker, who were both well-versed in their language and had cultural sensitivity and awareness to lead the interviews and focus group discussions. The researchers compensated the Rohingya women's time with raw basic food products, provided temporary child-care assistance while they were participating in the research, and transportation to the research location from their home. Informants of this research who had physical and/or mental health, or even other needs, were provided with assistance by RWDN and their network of relevant agencies or individuals (i.e., community leaders, activists). --- Findings The findings presented in the following four sections were selected based on the perspectives of the informants in this research. Only selected quotes and findings relevant to this article were chosen to be presented thematically here. Section 7.1 presents the experiences of women and girls being trafficked into the country and ending up in abusive marriages. Section 7.2 describes their experiences (including fears) and the experiences of pregnant women in accessing public hospitals. Section 7.3 reports on the financial barriers that Rohingyas face in accessing public hospitals, the influence UNHCR cards have on their accessibility, the sufferings they face, including mental health issues, and the difficulties in affording care. Section 7.4 reports on medical social work services, which are not provided to refugees and asylum seekers except for those under 18 years old. --- Women and Girls' Experiences of Marriage and Domestic Violence in Malaysia Rohingya women and girls who fled from persecution and violence in Myanmar found themselves to be experiencing violence within their households in Malaysia. Out of all the refugees and asylum seekers in the country, Rohingyas had the most people facing domestic violence issues. "It (violence) is one of the highest, in terms of sexual, gender-based, most common is domestic violence. And in domestic violence, the ethnic group that is known to be having these issues are the Rohingyas." (Refugee organization officer) The informants of this research mentioned having witnessed other Rohingya women and girls experiencing violence and assault. Their experience of being raped, which resulted in them becoming pregnant, caused further issues in their marital life, especially when they were "married off" to Rohingya men without being informed of their pregnancy status. One of the informants reported having faced physical violence at the hands of her spouse. "I have witnessed with my own eyes that the women were raped very badly (before reaching Malaysia). After being in Malaysia, men buy those women. They didn't know they are pregnant. When they know that they are pregnant, they start to argue with each other." (Rohingya woman, 37 years old) "He abuses by words and says angrily, 'I cannot move because I don't have card, where can I go, how can I afford?' Sometimes I was beaten violently that my body swelled." (Rohingya woman, 25 years old) The interviews we conducted further confirmed that these women and girls were trafficked into the country and married off to local Rohingya men regardless of their age. The issue of child marriage is difficult to address because it became a norm, and many of the community members hold on to it, especially the older men. Although the community is reported to practice child marriages, the extent to which it is prevalent in Malaysia is not known. "And there is also the issue of child marriages. There is still that practice among Rohingyas that if you are 14, 15, no more education for you... marry you off." (Refugee organization officer) Additionally, according to the refugee organization officer, the most common forms of SGBV faced by Rohingya women and girls in Malaysia were marital rape, domestic violence, child rape and sexual assaults. The perpetrators were usually people whom they knew such as their husbands, community members, locals, employers and teachers. Cases of SGBV by teachers at refugee schools were increasingly reported in the community. Additionally, the activist informant, who had worked with refugees for more than five years, reported that child marriages and the scale of sexual violence endured by Rohingya women and girls were appalling. "A few thousand girls between the ages of 11 and 15 or 16 that have been subjected to that kind of horrific sexual violence that you and I can't even comprehend and then sold into marriages to Rohingya men, trapped forever in this institution, stupid institution called marriage, you know, having to endure sex at the whims and fancy of the husbands, the so-called 'husbands'." (Activist) Consequently, the mental health problems of Rohingya women and girls were further affected by the violence they faced from their husbands and other family members. "when we ask, not all of them, but some, are willing to share more, have mental disorders before they come into Malaysia. We also have some patients who have already developed schizophrenia" (Mental health service provider) --- Access to Public Hospitals for Refugees and Asylum Seekers Rohingyas accessed public hospitals that were close to where they lived and worked. In the hospital where this research was conducted, a high number of Rohingya patients were seen because it was located near a wholesale market where a high number of people from this ethnic group live. "They work under the owners of the 'pasar borong' (wholesale market), and they run the markets themselves. So, these ladies are the wives of those who are working there." (Medical officer) The registration and possession of a UNHCR card had a significant impact on Rohingyas' access to healthcare services, particularly in public hospitals. Among Rohingya women and girls, those who are pregnant and those with serious health conditions were the ones most at risk, as they faced barriers to receiving medical assistance. The majority of the challenges and fears associated with pregnancy and delivery were shared by the younger Rohingya informants in this research. The process of registering refugees and asylum seekers for admission and treatment in hospitals was reported to be challenging. One of the volunteer informants of this research felt resistance from a hospital's registration counter when she attempted to register one of the Rohingya women and children. "As soon as they know you are dealing with refugees, you get the same treatment the refugees being given." (Volunteer) Rohingya women, particularly those without a UNHCR card, reported they avoided visiting hospitals for fear of being arrested, detained or unable to receive treatment. "I heard if people don't have UN cards, they go to hospital GH and if the people from the hospitals know that they don't have cards, they called immigration and handed over them to immigration [officers] to arrest them." (Rohingya woman, 18 years old) Another informant reported that she knew a girl who was arrested and detained for five months in a detention centre after being discharged from hospital. "There are some people who were suffering from sickness, went to the hospital when they don't have card or money, they depend on the contribution from the people, and they go to the hospital and after getting admitted in hospital and cured, police go there and arrest (them) and send them to the camp (detention centre." (Rohingya woman, 25 years old) Apart from the non-UNHCR card-holding Rohingyas facing barriers to accessing public hospitals, women among them specifically faced challenges in delivering their babies. "Since we don't have (the UNHCR) cards, if we go to the general hospital to deliver the baby, the nurse will ask whether we have the card or not. If we cannot show them the card, they shout at us, and it is very difficult to make the book for the pregnancy and delivery." (Rohingya woman, 23 years old) Some Rohingyas also reported having fled halfway during treatment at the public hospital for fear of arrest and detention, in addition to fear of being asked to pay. An interviewed volunteer shared her experience of working with Rohingyas, who tended to flee due to their fear that the authorities will arrest and detain them for failing to pay their hospital bills. "And there are those that have successfully received healthcare assistance from the hospitals but what happens is that halfway through the procedures they run because they fear that they are unable to pay and the police will come after them. This happens in the Government hospitals. Many times, we have sort of raised funds, put them in, the procedures are half-way ongoing, and then they just remove all the machines and they just run." (Volunteer) --- Financial Barriers to Access Healthcare Services The inability of refugees and asylum seekers to pay for their hospital bills and medicines has emerged as one of the main barriers to their access to public hospitals in Malaysia. "I think they always indicate that they have difficulties accessing care, and one of the main reasons is also because of affordability of care. Many of course work with very minimal pay, when they have a medical issue, they try to self-medicate. And then only go to the hospital or seek care only when they can't take it anymore, when it gets serious. So most of them also have issue because of documentation." (Refugee organization officer) All Rohingya informants mentioned that they faced challenges accessing hospitals because they could not afford the cost of the treatment and medicine charged. The amount of money they earned was barely sufficient to run their family here in Malaysia and to remit some to their relatives in Myanmar. They said that when they were sick, they did not go to hospitals, but instead went to private clinics because it was more affordable. "With one person income, I have to pay for house rent, school for children, our food. So how can we go to the expensive hospital? Sometimes I go there when I can afford. Sometimes I have to send money to Akhyab (home), for my relatives. We don't have a single cent as our savings. We are somehow just surviving." (Rohingya woman, 37 years old) The cost that they bore depended on whether they were registered with UNHCR and possessed its card. Refugees reported facing fewer issues with regard to admission and payment compared with asylum seekers (non-UNCHR card holders). Therefore, noncard holders were more likely to avoid public hospitals when they had healthcare issues requiring medical attention, than card holders. "If we go to hospital, it is very difficult to get treatment. We go to hospital when we don't have any other way because of the [inability to earn/having little] income and house rent [that we have to pay], (informant implying that they don't have any money left end of the month after paying for all the expenses to go to the hospital). If we go to the hospital, we are being charged too much money. We go to the hospital when we have no other way, since we don't have UN card." (Rohingya woman, 27 years old) Rohingyas who were connected with any NGOs or Malaysian (local) individuals who could assist them with admission and could guarantee that their bills would be paid, could sometimes gain better access to a public hospital than the rest of the population. "Again, they will also be charged on the foreigner's rate. No matter where they go, it will be the foreigner's rate (higher charge)....And they can't afford it. Due to this, most of them stopped going to seek (for) doctor's assistance. A lot of them have died. And because they just can't afford. I mean cancer patients, kidney failure, liver failure, they can't get assistance, they just die. I don't know how to put it, but they just die. I just know a lady's husband who just died because he couldn't afford." (Volunteer) Rohingya women faced challenges, especially when they delivered their babies, as the cost of childbirth in public hospitals was unaffordable for them. The cost increased if they needed any surgical procedure, suffered any complications, or if the baby needed further treatment and care. "And we have to save RM5000 to RM10,000 for surgery (caesarean section). If caesarean, we have to pay RM10,000 and if without caesarean, we have to pay RM5000. So feel very bad. Even (when) I am able to give birth to the children, I am not giving birth because of that because the charges are high." (Rohingya woman, 45 years old) The payment for childbirth for Rohingyas with UNHCR cards was less of an issue compared with those without the UNHCR cards. "If you come in labour, to pay as a foreigner is about, the deposit is about almost 3-4 thousand, so it is a lot of money." (Medical officer) The hospital's billing department staff and nurses reported to ensure that the payment for delivering babies was completed before both the mother and baby were discharged from the hospital. Upon delivering her baby, one informant recalled that she was asked to pay her hospital bill before she was allowed to see her baby. "When they called my husband, they told him to bring the money. So, if we ask, 'How much money do we need to pay?', they ask 1000, 1500. If we can't pay and if we want to see the baby, they say, 'You can't see the baby unless you can pay'." (Rohingya woman, 33 years old) Rohingya informants reported that they had been threatened to be handed over to the authorities if they failed to settle their hospital bills. A volunteer informant shared her experience of knowing Rohingya women who were unable to afford their delivery at public hospitals. "It is sad to know that you are expecting and you can't do anything to keep a healthy baby, when you deliver, you can't afford to have a delivery, so you opt to delivering at home and then when they do it at home, these aunties don't know what they are doing, but they still do it, and then they get the mothers get infected, the child gets infected, and the child suffers, the mothers suffer. I've seen these things happen. It is actually quite depressing." (Volunteer) The affordability of obtaining medical treatment was also mentioned to be an issue for Rohingyas seeking treatment for mental illness. The high charges for consultation and medicines, in addition to the lower quantity of medicines given, required patients to visit doctors more frequently, which increased the drop-out rates of Rohingya patients with mental illness. "Yes, this is why usually they will default treatment. Especially for the non-documented, the non-UN card holders will need to pay RM120 for each consultation, and they are only given five days medication every time. For example, you know schizophrenia, these are lifelong disorder, they have to take the medicine no matter how. If you give them (medicines) for only five days, they might go for the second time, third time, they might not have the money to go for the fourth time. They will discontinue the treatment saying the symptoms is likely to be in control, and later you will see the patient relapse." (Mental health service provider) Apart from medical bills for themselves and their family members, Rohingya women informants also reported facing financial challenges affecting other aspects of their lives, as most were not earning an income. They undertook unpaid housework and depended on their husbands or other male family members for money most of the time. However, when any Rohingya patient faced challenges in paying their hospital fee, the community members reported that they banded together to help the patient. "Usually in their community, if one person, let's say the lady who is pregnant, needs to go to hospital, they pool their funds. Yes, they pool their funds. So, sometimes they manage to pay. For cases they do not manage to pay, I have not come across any so far." (Medical officer) Although some may have received financial support, others did not because most people could not afford to provide such support. On the other hand, Rohingya women received social support from their extended family and friends. Social support relieved them from sadness, and provided them with needed information to navigate through life, in addition to companionship. --- Medical Social Workers and Their Work with Refugees and Asylum Seekers Similar to other public hospitals in the country, the services that medical social workers provided in the hospital where this research was conducted were for Malaysian citizens. They provided services to a range of patients with social issues such as victims of domestic violence, abused mothers, and children who were in need. They also provided services to those in need of practical assistance, which were not limited to financial support to receive treatment and purchase medical equipment, discharge planning, institutional placements and tracking of family members. As patients required referrals for social work services from medical officers from clinical departments, social workers only knew of patients who were referred to them. Furthermore, they only accepted cases and provided intervention to those who were eligible according to the department's policy. The cases of patients in need of medical social work services were filtered for their eligibility to receive services, before in-depth evaluations were conducted. "If we were to follow our SOP (Standard Operating Procedure), we don't take in foreigners, including refugees. If we get referrals, we will explain that it is not in our SOP to provide services for them. But we will advice on what they can do. If its billing, then they can settle with the billing department, but we can verbally tell them which NGO they can go to if they can't settle." (Medical social worker) For practical assistance, which involves financial support and institutional placements, the medical social workers followed their SOP and rules set by the resource providers or donors. "As for non-citizens, for most of the cases, we will not provide any support for financial assistance and institutional placement. Because for financial support, we are bounded by agencies which provide those resources. They [resource providers] have made it clear that its only for citizens." (Head of Department, Medical social worker) Social workers reported that they frequently received calls from hospital wards for financial assistance for refugees and asylum seekers. One of the medical social workers related a case of a patient in need of financial support for psychiatric treatment, but the department was not able to provide any assistance for the patient. "There was a case where the doctor wants to do treatment for a psychiatric patient, the rate was quite high, so the doctor requested whether can go through JKSP (the Hospital's Social Work Department)....they should have checked directly with HASIL (The Billing Department)." (Medical social worker) When medical social workers were unable to provide their services for the refugees and asylum seekers, they suggested that their clients approach other NGOs that might be able to assist them. These NGOs were suggested based on their willingness to provide financial support for medical treatments to non-Malaysians. The medical social worker stated that: "We can only provide advice to them, we can't go in-depth. We can't open [the patient case] file." (Medical social worker) However, under certain circumstances, medical social workers provided their services to non-Malaysian citizens. They handled cases involving children who were under the age of 18 years. This was because those who are below 18 are protected under the Child Act. "If they are underage, we have to accept them regardless whether they are refugee or foreigner." (Medical social worker) --- Discussion and Recommendation This research explored the experiences of Rohingya women refugees and asylum seekers in accessing healthcare services, and the challenges social workers face in providing their services to these women. These women refugees and asylum seekers face numerous physical and mental health issues due to the challenges they faced in Myanmar, on their journey to, and while in, Malaysia [7,14,45,46]. Despite their needs, Rohingya women face barriers to accessing needed healthcare services in Malaysia due to insecure legal status and protection challenges related to security, livelihood, poverty, deprivation, and social exclusion [14]. Additionally, they also face social and cultural barriers [7]. The findings of this research add to the existing literature on the challenges Rohingya women face in Malaysia, and in accessing healthcare services in public hospitals. At the point of arrival in Malaysia, they have acute and immediate health needs due to the dangerous and difficult journey they endured and their predisposing illnesses [7]. Similar to past studies [10,11,47], these women reported violence and assault prior to their arrival and while living in Malaysia. This research found that their experience of domestic violence while in Malaysia could be linked to the trafficking of girls and women for marriage to Rohingya men in Malaysia. According to [48], married Rohingya girls have higher risk of gender-based violence (GBV), including IPV, and mental distress. Environmental stressors, especially unemployment and money problems, are key triggers for the abusive behaviour exhibited by Rohingya men [48]. These victims of intimate partner abuse often do not seek for help from others due to fearing shame, social stigma, concerns about confidentiality, concerns about legal documentation and language barriers [46]. Apart from these factors, the acceptability and normalization of intimate partner violence by men among Rohingyas discourages help-seeking behaviour among the women, and those who do want to seek help prefer informal support such as from family and religious leaders [47]. While older Rohingya women's attitudes are changing, married girls are still excluded from social and personal progress [48]. Involvement of married girls and women in a range of programmes which improve their social and personal functioning, such as IPV, mental health, and family planning, is highly

The internal conflict in Rakhine State, Myanmar over the last few decades has forced Rohingyas to flee to other countries, including Malaysia. However, the non-recognition of the status of refugees in Malaysia keeps Rohingyas as "people without documents" and without adequate protection, access to healthcare services, education, and employment. Women among these groups face different challenges and are at risk of numerous physical and mental health issues. Thus, this research attempted to understand the experiences of Rohingya women in Malaysia, particularly in accessing public hospitals. Focus group discussions and key-informant interview techniques were used to collect data. The transcripts were then analysed using the thematic analysis method. The research found that Rohingya women experience numerous challenges despite being on safer ground in Malaysia. Their experiences of marriage and domestic violence, access to public hospitals, financial barriers to healthcare services, and the services provided by medical social workers in the country to refugees and asylum seekers are presented. Non-recognition of the status of refugees in the country is one of the primary barriers to the allocation and provision of resources for refugees and asylum seekers. Due to structural barriers, medical social workers are unable to provide their services to this population. This article provides recommendations for social workers in Malaysia on how to overcome these challenges and work more effectively with refugees and asylum seekers.

and while living in Malaysia. This research found that their experience of domestic violence while in Malaysia could be linked to the trafficking of girls and women for marriage to Rohingya men in Malaysia. According to [48], married Rohingya girls have higher risk of gender-based violence (GBV), including IPV, and mental distress. Environmental stressors, especially unemployment and money problems, are key triggers for the abusive behaviour exhibited by Rohingya men [48]. These victims of intimate partner abuse often do not seek for help from others due to fearing shame, social stigma, concerns about confidentiality, concerns about legal documentation and language barriers [46]. Apart from these factors, the acceptability and normalization of intimate partner violence by men among Rohingyas discourages help-seeking behaviour among the women, and those who do want to seek help prefer informal support such as from family and religious leaders [47]. While older Rohingya women's attitudes are changing, married girls are still excluded from social and personal progress [48]. Involvement of married girls and women in a range of programmes which improve their social and personal functioning, such as IPV, mental health, and family planning, is highly pertinent for women's empowerment [48,49]. There is a need for a comprehensive healthcare service [7] for refugees and asylum seekers beyond the basic services they can currently access in Malaysia. This research found that the physical and mental healthcare needs of Rohingya women are not adequately being met due to multiple challenges that exist. Financial access to specialized healthcare services in public hospitals, especially during admission and treatment, is beyond the affordability range of Rohingya patients. Asylum seekers, more than Rohingya refugees, in this research, face a heavier burden of affording healthcare as they are not entitled to the refugee discount due to lack of the UNHCR card, similar to the findings by other studies conducted in Malaysia [7,14,50]. However, refugees also find the cost of healthcare services at government hospitals to be unaffordable despite the 50% discount on fees charged to non-citizens [14]. Rohingya asylum seekers and refugees are unable to afford healthcare services primarily due to their lack of legal status in Malaysia, which prevents them from engaging in formal employment [50]. These refugees and asylum seekers remain in low level and underpaid informal work [14] that generates low, insufficient income to afford living expenses, including healthcare services. They also face language barriers [51], which are worse for Rohingya women and girls who are often highly dependent on male family members and relatives [52]. While the UNHCR Malaysia provides payment for medical bills for some refugees, these are often difficult to access [53]. Similar to [7], this research found treatment and transportation costs influence refugees' and asylum seekers' access to healthcare services, especially those with chronic illnesses, who require longer term care and regular follow-up. Medical social workers are unable to extend their services to refugees and asylum seekers in Malaysia due to several barriers. Currently, the SOP of social workers in Malaysian public hospitals explicitly lists the services that should be accorded to Malaysian citizens and non-citizens [54,55]. These structural barriers that link healthcare eligibility to the legal status of refugees and asylum seekers hinder medical social workers from delivering their services [56]. Previous studies conducted in another refugee-hosting, non-signatory country to the 1951 Refugee Convention and its 1967 Protocol, Bangladesh, similarly found Rohingyas face legal and administrative barriers to accessing health and security [57,58]. Given these challenges, there is a need for a change in policy and funding in addition to inter-sectoral coordination for social workers to provide their services to refugee and asylum seeker women [59]. Social workers need to undertake national level and state level policy advocacy to address the issues and needs of asylum seekers and refugees [56] generally, in addition to women and girls specifically. This macro-level practice of social work is crucial in refugee and asylum seeker host countries such as Malaysia, which lack legal recognition of this group of people's rights. Additionally, there is a need for social workers to identify, understand, and prevent oppressiveness through awareness of and overcoming the "issues of control that seek to justify bureaucratic aims rather than enhancing human well-being", in line with antioppressive practice (AOP) [36]. If resources are in place to provide services for refugees, social workers need to understand the challenges that different groups of refugees and asylum seekers face, keeping in mind how their intersecting social identities, such as gender, age and religion affect their social functioning. They can also use the situation-focused approach to acquire a systemic understanding of the refugee situation in the country [60]. Apart from that, the position of refugee social workers, who have expertise in working with the refugee and asylum seeker population, needs to be created and placed in key government institutions where these groups of people frequent, such as hospitals, immigration departments, detention centres, and in areas or districts in the state where they are concentrated. Professionals with expertise and experience working with both the population and on gender issues need to provide capacity building, training, and/or mentorship for social workers. In addition, given the complexities within the Rohingya community, refugee social workers need to work with boys and men to address the prevalent issues of domestic and intimate partner violence. One of the limitations of this research is it did not go beyond gender binary in its analysis and reporting. The researchers did not explore the perspectives of men, non-binary Rohingya refugees and asylum seekers. Future research could focus on non-binary and transgender Rohingya refugees and asylum seekers, as there is a paucity of literature on these groups of people. The researchers also did not consider other intersectional social identities of Rohingya women, as it was beyond the scope of this research. Despite its limitations, the research provides an insight into the important challenges Rohingya women and girls face within their family, and in accessing healthcare services that they need. --- Reflexivity and Positionality In all social science research, a researcher's positionality influences the way they conduct the research. Therefore, this section will be helpful in understanding our positionality and the lens we used for this research. This research was conducted as part of the first author's PhD project. The author has worked as a gender researcher in an organization focused on food systems, and has completed some training on the subject. The second author was the project supervisor, who is also a professor of social work with more than 20 years of experience in their field. The first author initially had a brief engagement with Rohingya women refugees and asylum seekers in Selangor, Malaysia, through a volunteer worker, prior to designing and conducting this research. At the meeting, the experiences of the Rohingya community, namely, the challenges they faced in this country, were shared and further discussed. The author was in frequent communication with the volunteer and others, including a woman and a man from the community, who kept the researcher up-to-date on the assistance provided to and any new initiatives for Rohingyas. Most of the engagement was with women and girls, who were facing numerous challenges and lacked needed support. The first author is aware of the privilege he holds (and bias he may have) as a man and having citizenship of a country, while the women and girls interviewed for this project are of a different gender and are stateless. The author was able to be conscious of his thoughts and actions through reflection moments during coffee sessions with the two ethnic Rohingya volunteers, who shared their personal experiences growing up and living their lives as refugees. Additionally, the project supervisor, as a female social work professor, ensured that the research was sensitive to women and girls from the Rohingya community. --- Conclusions The systematic discrimination, statelessness and targeted violence aimed at Rohingyas over the past few decades in Rakhine State, Myanmar, triggered hundreds of thousands of Rohingyas to flee their country for other countries, including Malaysia. The trauma and suffering they endured in their home country during the conflict, during the journey to, and in the host country, are physically, emotionally and mentally agonizing. The protracted situation due to the non-recognition of their status in Malaysia further negatively impacts their state of being. Among this group of people, women and girl refugees are exposed to physical and mental health issues due to abuse and exploitation, including forced marriages. Therefore, this research attempted to understand the experiences and challenges that Rohingya women in Malaysia face with regard to their health, their experiences in accessing healthcare services, and the ways that medical social workers can enable this group of people to better access healthcare services in Malaysian public hospitals. The research found that Rohingya women were, and continue to face, numerous challenges which impinge on their physical and mental health. Despite their coping strategies, if policy makers, social workers, researchers and other stakeholders do not take measures at different institutional levels, Rohingya women in Malaysia will not be able to have better access to health, and consequently a better life. --- Funding: The cost to conduct this study was funded by the authors as part of the first author's PhD thesis. The publication fee for this article was funded by the Faculty of Social Sciences and Humanities, Universiti Kebangsaan Malaysia. This information or content and conclusions are those of the author and should not be construed as the official position or policy of, nor should any endorsements be inferred by, the organizations the authors are affiliated with. --- Institutional Review Board Statement: The Human Research Ethics Committee of Universiti Sains Malaysia (HREC) (Code: USM/JEPeM/17020143) and the Medical Research and Ethics Committee (MREC) (ID: NMRR-17-3136-38699) approved this research. The first body was responsible for all studies conducted by the University's researchers while the second body was responsible for all studies conducted within the premises of the Ministry of Health, Malaysia. Informed Consent Statement: Informed consent was obtained from all participants involved in this study. --- Data Availability Statement: The data presented in this study are available on request from the corresponding author. The data are not publicly available to retain informant privacy. --- Conflicts of Interest: The authors declare no conflict of interest.

The internal conflict in Rakhine State, Myanmar over the last few decades has forced Rohingyas to flee to other countries, including Malaysia. However, the non-recognition of the status of refugees in Malaysia keeps Rohingyas as "people without documents" and without adequate protection, access to healthcare services, education, and employment. Women among these groups face different challenges and are at risk of numerous physical and mental health issues. Thus, this research attempted to understand the experiences of Rohingya women in Malaysia, particularly in accessing public hospitals. Focus group discussions and key-informant interview techniques were used to collect data. The transcripts were then analysed using the thematic analysis method. The research found that Rohingya women experience numerous challenges despite being on safer ground in Malaysia. Their experiences of marriage and domestic violence, access to public hospitals, financial barriers to healthcare services, and the services provided by medical social workers in the country to refugees and asylum seekers are presented. Non-recognition of the status of refugees in the country is one of the primary barriers to the allocation and provision of resources for refugees and asylum seekers. Due to structural barriers, medical social workers are unable to provide their services to this population. This article provides recommendations for social workers in Malaysia on how to overcome these challenges and work more effectively with refugees and asylum seekers.

Background Risky drinking habits, including heavy episodic drinking (HED), comprise a major burden for public health [1]. The definitions of HED vary, but it is usually defined as drinking 4-6+ alcohol units in one occasion. HED is associated with short and long term health problems, and social and socioeconomic harms to self, others and for society at large [1]. HED is common in Europe, but it usually decreases with age, notably after a person enters their thirties [2]. One of the challenges is to identify those who continue this detrimental health habit into later adulthood. Thorough understanding of the complex developmental processes related to alcohol use requires deeper knowledge on the individual, contextual and life course aspects, and especially their interrelations that influence this development [3]. Combining these aspects is not common in previous empirical studies on alcohol use. Bronfenbrenner's bioecological theory of human development and the Process-Person-Context-Time (PPCT) model [4] provides a useful framework for understanding the complex relations between an individual, their environment, and the passage of time in influencing drinking habits. --- Process In the PPCT model the first concept 'process' refers to dynamic interplay between the individual and their environment (including other persons, objects and symbols) through which development occurs. Proximal processes refer to associations between an individual and their immediate environment. This immediate environment has been referred to as the'microsystem'. Social relationships represent a domain where proximal processes occur between an individual and other persons. In this study we focus on processes occurring in adolescence; a time when interrelations between an individual and their family, peers, and school, constitute important processes that effect the person's development. Adolescence is often a time of rapid changes in relationships; child-parent relationships are often renegotiated [5] and the importance of relationships in other growth milieus strengthen. For example, life style choices and decisions to engage in risky behaviours such as alcohol use may be more influenced by peer relationships [6]. However, despite the increasing influence of other factors, family relationships retain an important role in the lives of adolescents [5,7]. Many studies have examined the effects of family and peer relationships on adolescent alcohol use. Several previous studies have concluded that good adolescent family relationships protect against adolescent heavy alcohol use [8][9][10]. However, a systematic review found mixed evidence for parent-child relationships and adolescent alcohol use, and called for more detailed and versatile information on alcohol use [11]. In addition, adolescents who are more peer-oriented tend to use more alcohol [12,13], while poor school bonding and environment (e.g. a student's low attachment to teachers, poor school satisfaction) also increase the risk for alcohol use and drunkenness [10,14]. Although there is some lack of clarity in the association between adolescent social relationships and adolescent alcohol use, the evidence for these concurrent associations is stronger than that for associations that extend beyond adolescence. Regarding associations between adolescent social relationships and later adulthood alcohol use the evidence is mixed. Some studies have found associations between poor family relationships (e.g. less harmonious relationships, family conflict) in adolescence and harmful alcohol use in later adulthood [15,16], while others have not [17,18] or have only found associations in either gender or only at some ages [19][20][21][22]. Findings regarding peer relationships (e.g. low peer acceptance) are equally mixed [18][19][20]23]. Evidence suggesting associations between adolescent school environment and adult alcohol use is scarce [21]. --- Person In Bronfenbrenner's PPCT model the 'person' component refers to an individual's characteristics (e.g. age, sex, heavy episodic drinking). These are the characteristics that a person brings with them to any social situation [4,24]. Men tend to drink alcohol in larger amounts and more often than women [25]. Whether the influence of the risk factors on drinking also differs between sexes is not so evident. Sex differences related to the role of family and other social relations are especially under studied [26]. Sex could also be seen as a more contextual measure (i.e. gender), but in this study we follow Bronfenbrenner's view of sex as an individual level measure. --- Context In the PPCT model 'context' describes the environment of the person, whereas 'process' describes the way the person and context are interrelated. In the PPCT model the 'context' can be divided into four different systems. The microsystem refers to the immediate environment of a person. For example in adolescence, family, peer group, school and hobbies form important settings in the microsystem. The mesosystem refers to interrelations between different settings within the microsystem. It has been suggested that constellations of relationships in different settings may matter more than relationships in any specific setting; for example family and peer influences can both contradict and reinforce each other [6], such as where good social relationships in the family setting (e.g. good family cohesion) can buffer against negative peer context influences (e.g. friends drinking alcohol) on alcohol use [27,28]. In previous studies on social relationships in various settings and alcohol use, it has been common to examine the effect of one setting at a time, rather than the interplay between different settings. One study that did examine several different settings simultaneously found family setting to be a more important predictor of HED at age 33 than risky peer setting or school setting [21]. The exosystem refers to an outer layer of the context, not where the individual of interest is directly situated, but which has an indirect influence on their development (e.g. parent's or spouse's workplace). Finally, the macrosystem is the outermost layer, which refers to the overarching pattern of characteristics of a group, whose members share resources, hazards, lifestyles, opportunity structures, life course options and patterns of social interchange [24,29] (e.g. government policies, economic circumstances, cultural expectations). In this study we consider the macrosystem in the form of two different countries: Finland and Sweden. While both countries are Nordic welfare states, some differences have occurred especially regarding differences in alcohol policies and drinking habits. HED is a common drinking habit in both countries, although more so in Finland than Sweden [1]. Alcohol use has increased in both countries from the 1960s onwards until very recently, but more rapidly in Finland [30]. Alcohol policies (regulation of sales, alcohol taxation, retail policy etc.) have been rather strict in both countries. Some macro level differences have also been observed regarding family and school. Regarding family policies Sweden has been a forerunner in introducing parental leave [31], and the culture and policies in general have promoted more equality between women and men than in Finland. No major differences in child-rearing practices have been observed between the countries, but it has been stated that based on "equality ideology", even young children are treated as equal partners to other family members in Sweden [32] and in comparison, the Finnish interactional structure is less equal [33]. School systems in Finland and Sweden both strived for social equality in the 1980s, but the focus is now more on individuality [34]. --- Time The final concept 'time' adds temporal aspects to the model, which are central to the idea of development. Time can refer to the chronological age of an individual, but also to historical time. In this study, the focus is on the former in terms of individual development through different life stages. Adolescence is a crucial time regarding alcohol use, because it is often initiated during this time and the drinking habits adopted tend to persist into adulthood [35]. Although the associations between social relationships (i.e. social support, networks and integration) and health behaviours are well established, for a long time there has been a distinction between studies that focus on either adolescence or adulthood [6], with the life course perspective generally being overlooked. This point was noted in 2010 by Umberson et al. and studies have since examined associations between adolescent social relationships and later alcohol use, however few researchers have examined these issues with long follow-up times spanning over several life stages. One of the few studies examining the unique and combined effects of adolescent family, peer, and school settings on HED in adulthood found that family relationships were associated with HED at age 33 but not at age 21, concluding that some of the contextual risk factors may gain their salience as people age [21]. As such, it is important to study alcohol use in several different life stages. Many previous studies applying the PPCT model have not considered all aspects of the model nor analyzed in detail the relations between those aspects [36]. The PPCT model emphasizes the importance of the interrelations between its four constituent concepts. For example, personal and contextual characteristics likely interact, but the details of these interactions may be blurred. It might be that on the macro level, the more drunkenness-oriented drinking culture in Finland is more detrimental to men, but how the supportive proximal processes may protect from this macro feature is unclear. --- Aims Using Bronfenbrenner's PPCT model as a framework, this study examines how adolescent proximal processes, namely social relationships in various settings and the interactions between them (mesosystem) are associated with heavy episodic drinking over the life course in women and men, and the role of the macrosystem (country) in these associations. The more specific study questions are as follows: 1, the attrition rate in the Swedish cohort has been extremely low, while somewhat higher in the Finnish sample. At the age 42 follow-up of the Finnish cohort, the non-participants were more frequently men and had poorer school performance at age 16 compared with the continuing participants. Non-participation at age 42 in the Finnish cohort was also associated with frequency of HED at age 16, but this difference can be explained by male preponderance among the non-participants. --- Measures --- Social relationships at age 16 in family, peer group and school settings For social relationship variables, indicators were selected that are most similar between the two cohorts (scales were from 1 to 5, if not otherwise stated). Transient conflicts with family during adolescence can be common, so using various measures of family relationships provides a more stable assessment of family relationships [37]. Proximal processes related to family setting consisted of four different indicators: 1) poor relationship with mother, 2) poor relationship with father, 3) poor general family atmosphere, and 4) not spending much time with family. In the Finnish cohort, the mother and father relationship variables (sum scales, range 3-15) were formed by three statements for each parent: i.'my mother/father is close to me' (reversed), ii. 'I often argue with my mother/father', and iii. 'I feel my mother/father understands me' (reversed) (all statements: 'totally disagree' to 'totally agree'). In the Swedish cohort, mother and father relationships were measured as one question each: 'how is your contact with your mother/father?' ('very good' to'very poor/no contact'). General family atmosphere was measured with response to one statement/question: 'I feel the atmosphere at home is good' (reversed) ('totally disagree' to 'totally agree') (Finnish cohort) / 'how do you enjoy/like being at home?' ('very well' to'very poorly') (Swedish cohort). Time spent with family was also measured with response to one statement/question: 'I spend most of my spare time with my family' (reversed) ('totally disagree' to 'totally agree') (Finnish cohort) / 'are you most often at home or away in the evenings?' ('almost always at home' to 'almost always away') (Swedish cohort). The mean of these four family indicators was calculated to form a general poor family relationship index. Relationships with peers and school settings were each assessed with a single question. Peer contact was measured with a question regarding time spent with friends: 'how often do you spend time with friends outside school hours?' ('daily' to'seldom') (Finnish cohort) / 'are you most often by yourself or with friends on spare time?' ('almost always with friends' to 'almost always alone') (Swedish cohort). A lot of time spent with peers was classified as daily peer contact. Relations with classmates were assessed with a question: 'how do you get along with your classmates?' (4-point scale, 'get along with everybody' to 'nobody') (Finnish cohort) / 'do you like being with your classmates?' ('very well' to'very poorly') (Swedish cohort). All social relationship variables (if not originally) were recoded to follow a scale from 1 (best/lowest risk) to 5 (worst/highest risk). For the interaction analyses, the variables were dichotomized as close to the upper quartile (75%) as possible. --- Heavy episodic drinking The United Nations define HED as having at least 60 g of pure alcohol (5 units in Finland and Sweden) on at least one occasion in the past 30 days [38] and this definition was applied as accurately as possible in the present study. HED was dichotomized (yes/no). In the Finnish cohort at age 22, the participants were defined as 'heavy episodic drinkers' if they reported heavy drunkenness at least monthly and 'not heavy episodic drinkers' if they reported heavy drunkenness less often (on a scale of 'once a week or more often', about 1-2 times a month', more seldom' or 'never'). At ages 32 and 42, HED was measured with a question 'how often do you have six or more drinks in a row?' from the Alcohol Use Disorders Identification Test (AUDIT) ('never', 'less than monthly', monthly', 'weekly' and 'daily or almost daily') [39]. Those who reported having six or more drinks in a row at least monthly were classified as 'heavy episodic drinkers', while those drinking less often (or not at all) were classified as 'not heavy episodic drinkers'. In the Swedish cohort the respondents reported the frequency of drinking occasions per beverage type ('every or every other day', '1-2 times a week', 'a couple of times a month','more seldom' and 'I do not drink this type of beverage') and average intake of beer (number of bottles), wine (number of glasses), and strong beverages such as spirits (number of drinks) on a typical occasion. Respondents who reported drinking on average <unk> 5 bottles of beer, <unk> 5 glasses of wine, or half a bottle (37 cl) or more of strong beverages in one session and drinking at least a couple of times monthly were classified as 'heavy episodic drinkers'. Respondents who reported drinking the aforementioned amounts more seldom (or not at all) were classified as 'not heavy episodic drinkers'. --- Country Country variable was indicated by the cohort source i.e. Finland (TAM) or Sweden (NoSCo). --- Control variables In the Finnish study, HED at age 16 (baseline) was defined as being drunk at least four times during the school term (on average once a month) and not HED if drunkenness occurred 0-3 times. In the Swedish study, HED at age 16 was measured as at the other ages. For both cohorts, parental socioeconomic position (SEP) at age 16 was classified as 'non-manual' vs.'manual' employment based on father's occupation [40,41]. --- Statistical analyses Bivariate and multiple logistic regression analyses with odds ratios (ORs) and 95% confidence intervals (CI) were used to examine how social relationships at age 16 are associated with HED at ages 22/21, 32/30 and 42/43. Analyses were performed for each microsystem setting (family, peer group, school) separately first with no adjustments (Model 1), and second controlling for HED and parental SEP at age 16 (Model 2). These analyses were stratified by country and sex. The cohorts were then pooled and interactions between the three settings in adolescence (family, peer group, and school) as well as the macrosystem (Finland vs. Sweden) in relation to associations with HED were assessed with additive interaction analyses in multiple logistic regression models. The additive interaction analyses were performed with six combined variables (family/peers, family/school, peers/school, family/country, peers/country, school/country) (see classification in Table 4). Relative excess risk due to interaction (RERI) is a measure of additive interaction [42]. RERI = 0 indicates the absence of an interaction or mere additivity; RERI > 0 indicates a positive interaction or an effect that is more than additivity (i.e. having the two risk factors constitutes a more pronounced risk for an outcome than what would have been expected by just summing the separate risks of the two factors), while RERI <unk> 0 indicates a negative interaction or less than additivity. The 95% CIs for the RERIs were calculated as suggested by Knol and VanderWeele [43]. An interval not including 0 indicates statistical significance. All analyses were made separately for women and men. IBM SPSS Statistics 22 [44] was used for data analyses. --- Results Characteristics of the participants are summarised in Table 2. Men tended to be heavy episodic drinkers more often than women in both countries and Swedish women had a lower level of HED also compared to Finnish women, e.g. HED being twice as likely in Finnish women in the two last follow-ups. The logistic regression analyses for the HED outcomes for women and men, in Finland and Sweden are presented in Table 3. For Finnish women in the unadjusted model, higher levels of poor family relationships and peer contact were associated with an increased likelihood of HED in all ages, while associations with poor relationships with classmates were not significant. After controlling for HED at age 16 and parental SEP, poorer family relationships were associated with an increased likelihood of HED at ages 22 and 32 in Finnish women. In addition, a higher level of peer contact was associated with an increased likelihood of HED at ages 32 and 42. For Swedish women, higher levels of poor family relationships and poor relationships with classmates were consistently associated with an increased likelihood of HED at all ages in the unadjusted analyses, while only poorer relationships with classmates were associated with an increased likelihood of HED in the adjusted model and only at age 30. In the unadjusted model for Finnish men, higher levels of poor relationships with family and peer contact were associated with an increased likelihood of HED at ages 22 and 32, but poor relationships with classmates were not. With regards to HED at age 42, only the association regarding poor family relationships was significant. After adjustments, the associations with HED at ages 22 and 32 remained, while the association with poorer family relationships and an increased likelihood of HED at age 42 became non-significant. In the unadjusted model for Swedish men, all variables in adolescence were associated with an increased likelihood of HED at age 21. At age 30, only the association with higher levels of poor family relationships during adolescence remained significant. At age 43, both higher levels of poor family and classmate relationships, but not higher level of peer contact, were associated with an increased likelihood of HED. In the adjusted model, all associations except that for poor relationships with classmates remained significant at age 21. The association with poor family relationships remained at age 30, while none persisted at age 43. Interactions were analysed using the concept of additive interactions in multiple logistic regression analyses (Table 4). RERIs with 95% CI were calculated for all outcomes. The results indicate a synergistic interaction between daily peer contact and country for HED at age 42/ 43 in women (Fig. 1), whereby those women who spent a lot of time with peers and were living in Finland had an increased risk for HED. --- Discussion Applying bioecological theory and the PPCT model, this study examined adolescent social relationships as proximal processes in various contexts and their association with heavy episodic drinking from youth to midlife. We found poor adolescent family relationships to be associated with HED for participants during their twenties and thirties, but after taking into account HED and parental SEP at age 16, not once they reached their forties. Higher level of peer contact was also associated with later HED, while poor relationships with classmates were associated to a lesser extent. We did not find support for adolescent social relationships in one setting, neither family, peer group nor school, to be jointly associated with other settings on drinking behaviour in later life stages. Regarding the macro level, Swedish women were notably different from the Swedish men and Finnish men and women, demonstrating a lower level of HED, as well as a lack of associations between adolescent family and peer relations and subsequent drinking. This study combined contextual and life course perspectives, which has not been common in previous research on social relationships and alcohol use. Unlike most previous studies using Bronfenbrenner's PPCT model, we applied all four concepts: Process, Person, Context and Time. In general, we found support for the relevance of all four concepts of this theoretical framework. --- Proximal processes As adolescents make the transition to becoming independent, conflicts in the family may increase, especially if the process involves opposing parental control and trust, such as in the practice of engaging in alcohol use [45]. The direction of associations is likely complex; drinking may both increase family conflicts, while also being a way to relieve stress resulting from poor family relationships. The strongest evidence to support the association between poor parent-child relationships and subsequent alcohol use has been found in genetically informative multilevel studies on externalizing symptoms (e.g. substance use, conduct disorders) [46,47]. The results of the present study support previous findings. Poor family relationships in adolescence were associated with later life HED in Finnish women and men, and Swedish men, after taking into account HED in adolescence. Nonetheless, this association was only relevant regarding more proximal life stages (i.e. ages 22/21 and 32/30, but not 42/43). Relationships with parents typically form the primary social environment and basic foundation for a person's future development. Parents influence their children via numerous ways, such as their own alcohol use, attitudes, control, and support. Open and constructive communication about alcohol use is easier if the relationships between the adolescent and their parents are good [48]. In adolescence, peer relationships can have positive or negative effects on an adolescent's behaviour; either by encouraging or protecting from risky behaviours, depending on the norms and values among peers. Spending a lot of time with peers during adolescence was associated with later HED in Finnish women (ages 32 and 42, [age 22 only in unadjusted model]) and men (ages 22 and 32), and in Swedish men (age 22) in the adjusted models. Although social contacts are generally good for adolescent development, it may present a risk for later drinking behaviour. This finding highlights the complexity of the role of adolescent social relationships in later drinking behaviour [26,49]. Personality characteristics may in part explain this association, e.g. extroversion has been found to be associated with increased alcohol use/problems [50]. Unfortunately, we could not take into account whether the participants' peers were heavy episodic drinkers, which would likely have shed more light on these findings. However, we have taken into account the person's own HED, which usually correlates highly with peers' drinking habits [27]. A British cohort study that examined many risk factors for alcohol problems found peer group deviance to be associated with alcohol problems at age 20 [16]. We found poor relationships with classmates to be a risk for later HED only for Swedish women (all ages), and after taking into account age 16 conditions, only at age 30. Previous studies have found school related factors to be associated with drinking. However, most of these studies have examined concurrent drinking during adolescence and it could be that the association regarding school does not persist to later life. In the transition from compulsory school (age 16) to further education the relationships with classmates often change. So perhaps the influence of these relationships is not as important as for those that exist outside school. Our results are in line with a study that examined adolescent family, peer and school settings and HED at ages 21 and 33 [21], which did not find school factors to predict subsequent HED. However, unlike some previous studies --- Person Sex differences in heavy episodic drinking were as expected; men tended to be heavy episodic drinkers more often than women in both countries. The Swedish women tended to drink less heavily than all three of the other groups. Later life HED was associated with poor family relationships and frequent peer contact for both Finnish and Swedish men even after adjustments. The associations among Finnish women resembled those found in men, but the same pattern was not observed among Swedish women. In contrast with the other groups, for Swedish women poor relationships with classmates was the only variable of interest associated with later life HED. This could be due to the low prevalence of heavy episodic drinkers among this group. The level of HED in Swedish women is likely actually lower than in other groups, but it may also be that Swedish women were more likely to mix different beverages and our measure does not capture their HED as effectively as with other groups. In the current study we followed Bronfenbrenner's classification of sex as an indicator of an individual level characteristic. However, future studies with a focus on theory development could benefit from a broader view on gender indicating social order and contextuality, especially regarding studies on social relationships. --- Context The results suggest that the most relevant settings in the microsystem are family and peer settings. The mesosystem refers to interactions between these settings; however we did not find any significant interactions between settings regarding subsequent HED. Previous studies have found a good family setting to protect from the detrimental effects of a risky peer setting on alcohol use [27,28]. However, these studies have examined concurrent alcohol use in adolescence. In this study we did not find these kinds of protective elements regarding later drinking in adulthood, and it may be that this protective effect does not extend into later life stages. The role of the macro level has been less studied regarding associations between social relationships and drinking. In this study we noted differences between the countries in the associations between adolescent social Fig. 1 An additive interaction between peer contact and country for heavy episodic drinking (HED) at age 42/43 in women relationships and later life HED. In the additive interaction analyses, between country and both family relationships and peer contact, the odds ratios were generally higher for men in Finland compared to men in Sweden, regardless of the presence of the other risk factors. In adjusted interaction analyses the only significant RERI was found between peer contact and country regarding HED at age 42/43 in women. It was observed that those exposed to both daily peer contact and a more drunkenness-oriented culture (Finland) had an increased risk of drinking heavily in midlife. Along with family, peers form a social context through which an adolescent is exposed to cultural norms and influences [27]. --- Time Most previous studies have examined the association between social relationships and concurrent alcohol use in adolescence or examined these associations in later life stages but using cross-sectional designs. This study demonstrates that these associations extend beyond adolescence. We found more associations between adolescent social relationships and more proximal HED (at age 22/ 21) than at later life stages, with the number of significant associations decreasing with age. This finding is consistent with the proximity hypothesis, which suggests that more proximal exposures (i.e. those occurring closer to the time of the outcome) are more important than distal exposures [51]. Conversely, the association between frequent peer contact and HED emerged only at ages 32 and 42 in Finnish women, which is consistent with an alternative hypothesis from Lee et al., which states that some effects become more pronounced as people age [21]. It is well known that late adolescence and youth are typically times of peak alcohol consumption, and after this HED usually decreases especially in women, possibly due to parenting responsibilities [52]. It may be that those women, who still drink heavily in their forties, deviate in other ways from those who follow the typical pattern and could also have other additional risk factors for drinking. --- Methodological considerations The main strengths of this study are the long follow-up times and practically full participation rate at baseline in both cohorts. In the Swedish cohort the participation rate has continued to be extraordinarily high. Conversely, there was some attrition related to male sex and poor school performance at age 16 in the Finnish cohort's follow-ups. Although there was attrition related to HED at age 16 in the Finnish cohort, it was explained by male preponderance among the non-participants at age 42, which suggests that the results were not highly biased regarding heavy episodic drinking. Another particular advantage of this study is the possibility to examine the study questions with two separate but relatively similar datasets. While the measures used had been developed independently in the two cohorts they were harmonised as well as possible for the purposes of the present study. Still, the measures did differ between countries. The measure of HED in the Swedish cohort does not consider persons who combine different beverages as heavy episodic drinkers, if the amounts within one beverage type did not exceed the limit used in this study. Therefore the prevalence of HED in this cohort may be underestimated; yet the frequencies observed are rather similar to those found in other population based Swedish studies that measure HED more specifically [53]. In addition the measurement of HED changed in the Finnish study (i.e. 'perceived drunkenness' vs. 'having <unk>6 drinks in one occasion'). Perceived drunkenness has been shown to correlate with consumption of six units of alcohol in adolescence [54], suggesting that the HED measure generally captures the same drinking habit in all waves. Nonetheless, it should also be acknowledged that at age 22 the wording of the questionnaire in the Finnish cohort does refer to heavy drunkenness, so it may not capture milder drunkenness at that age, and this could explain the lower prevalence of HED in Finnish compared to Swedish men at this life stage (29.6 vs. 43.5%). There are also some issues considering the measurement of social relationships that should be taken into account. The measures in different settings of the microsystem may have measured different aspects of social relationships. In family and school settings, the quality of the relationships was measured, while in the peer setting the time/frequency spent among peers was the focus. The measurement of parental SEP in adolescence was based on the father's occupation, and if not available on the mother's. This decision to base SEP mainly on father's occupation was made because of a rather large proportion of missing information on mother's occupation in the Finnish cohort. It would be better to measure parental occupation based on both parents, but our procedure has been quite common in the 1980s. We did not specifically examine the possible mechanisms behind the associations between adolescent social relationships and later HED [55]. For example, it may be that patterns and skills of social interaction are adopted in childhood and adolescence, which are further reflected in the social relationships experienced in adulthood. Difficulties in social relationships in adulthood could also be associated with concurrent drinking behaviours [6]. The participants lived in one Finnish or one Swedish city during adolescence. In Finland some regional differences in adolescents' alcohol use have been found previously [56], and therefore the results are perhaps not generalizable to the whole Finnish population. However, the results are likely generalizable to Finnish urban adolescents of that time. The Swedish cohort has been shown to be comparable to the country as a whole with regard to socio-demographic and socio-economic factors as well as health status and health behaviors [57,58]. Isolating the causes of historical change or different aspects of the macro level (e.g. drinking culture vs. alcohol policy) in alcohol use is very complex given the net of likely multiple influences that are present simultaneously. More detailed information on macro level differences and the distinction between age, historical period, and cohort effects, would benefit further studies to trace the phenomena in greater detail. --- Conclusions Guided by bioecological theory and the Process-Person-Context-Time model, this study found proximal processes of poor adolescent family relationships and frequent peer contact to be associated with an increased likelihood of heavy episodic drinking well into adulthood. We did not find long-lasting protective or risk-raising effects between different settings (mesosystem). These results suggest the importance of early interventions and promotion of positive social relationships in adolescence. It is important to apply a holistic perspective such as the PPCT model, acknowledging the individual, contextual and temporal aspects when examining alcohol use and planning preventive actions and treatment. This requires further development of theories combining these perspectives, as well as new empirical data. Theoretical development is also needed in relation to sex/gender and incorporating wider aspects of gender order into the PPCT model. --- Availability of data and materials The datasets generated during and/or analysed during the current study are not publicly available due to legal restrictions (Finnish Personal Data Act 523/1999, The Swedish Data Protection Act 1998:204) and the nature of the data (individual level data). Data are available upon request. Data requests for the Finnish data are reviewed in National Institute for Health and Welfare Finland for compliance with the original research purposes of the study project. Requests are subject to legal restrictions of the Finnish Personal Data Act. Requests for the data may be sent to: Information Resource Services Unit/Arto Vuori, National Institute for Health and Welfare, P.O. Box 30, 00271 Helsinki, Finland (contact: Arto Vuori, [email protected]). Swedish data is available upon request from the Principal Investigator Anne Hammarström ([email protected]), pending ethical approval. --- Authors' contributions AH, TH and OK participated in the data collection. NB performed all the statistical analyses and NB, OK, CGB, TH, TL and AH interpreted the analyses. NB drafted the manuscript, which was revised by the rest of the authors (OK, CGB, TH, TL, AH). All authors read and approved the final manuscript. --- Ethics approval and consent to participate For the Finnish study (TAM), the study protocol was approved by the Ethics Committee of the National Institute for Health and Welfare (formerly National Public Health Institute). For the Swedish study (NoSCo), study protocol was approved by the Regional Ethics Vetting Board in Ume<unk>. Participants were informed of the purposes of the study and that participation was voluntary. They were requested to indicate their consent by answering the survey questionnaire. --- Consent for publication Not applicable. --- Competing interests The authors declare that they have no competing interests. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: Applying the Process-Person-Context-Time (PPCT) model of the bioecological theory, this study considers whether proximal processes between the individual and the microsystem (social relationships within family, peer group and school) during adolescence are associated with heavy episodic drinking (HED), from youth to midlife, and whether the macro level context (country) plays a role in these associations. Methods: Participants of two prospective cohort studies from Finland and Sweden, recruited in 1983/1981 at age 16 (n = 2194/1080), were followed-up until their forties using postal questionnaires. Logistic regression analysis was used to examine associations between social relationships at age 16 and HED (at least monthly intoxication or having six or more units of alcohol in one occasion) at ages 22/21, 32/30 and 42/43. Additive interactions between microsystem settings, as well as between settings and country, were also considered. Results: Consistent with the PPCT model, we found individual, contextual and temporal aspects to be associated with drinking habits. Higher levels of poor family relationships were associated with an increased likelihood of HED (ages 22/ 21 and 32/30) in both Finnish women and men and Swedish men. Higher levels of peer contact were associated with an increased likelihood of HED in both Finnish women (ages 32 and 42) and men (ages 22 and 32), and Swedish men (age 21). In contrast with the other groups, poorer relationships with classmates were associated with an increased likelihood of HED (age 30) for Swedish women only. For women, the combined effect of having both daily peer contact and living in Finland for HED at age 42/43 was statistically distinguishable from a pure additive effect. Conclusions: Micro and to a lesser extent macro level contexts are associated with heavy episodic drinking well into adulthood. The most relevant processes in the adolescent microsystem occur in family and peer settings. However, long-lasting protective or risk-raising effects between different settings and later HED were not found. Promoting good relationships across different contexts during adolescence may reduce the incidence of HED in adulthood.

TH and OK participated in the data collection. NB performed all the statistical analyses and NB, OK, CGB, TH, TL and AH interpreted the analyses. NB drafted the manuscript, which was revised by the rest of the authors (OK, CGB, TH, TL, AH). All authors read and approved the final manuscript. --- Ethics approval and consent to participate For the Finnish study (TAM), the study protocol was approved by the Ethics Committee of the National Institute for Health and Welfare (formerly National Public Health Institute). For the Swedish study (NoSCo), study protocol was approved by the Regional Ethics Vetting Board in Ume<unk>. Participants were informed of the purposes of the study and that participation was voluntary. They were requested to indicate their consent by answering the survey questionnaire. --- Consent for publication Not applicable. --- Competing interests The authors declare that they have no competing interests. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: Applying the Process-Person-Context-Time (PPCT) model of the bioecological theory, this study considers whether proximal processes between the individual and the microsystem (social relationships within family, peer group and school) during adolescence are associated with heavy episodic drinking (HED), from youth to midlife, and whether the macro level context (country) plays a role in these associations. Methods: Participants of two prospective cohort studies from Finland and Sweden, recruited in 1983/1981 at age 16 (n = 2194/1080), were followed-up until their forties using postal questionnaires. Logistic regression analysis was used to examine associations between social relationships at age 16 and HED (at least monthly intoxication or having six or more units of alcohol in one occasion) at ages 22/21, 32/30 and 42/43. Additive interactions between microsystem settings, as well as between settings and country, were also considered. Results: Consistent with the PPCT model, we found individual, contextual and temporal aspects to be associated with drinking habits. Higher levels of poor family relationships were associated with an increased likelihood of HED (ages 22/ 21 and 32/30) in both Finnish women and men and Swedish men. Higher levels of peer contact were associated with an increased likelihood of HED in both Finnish women (ages 32 and 42) and men (ages 22 and 32), and Swedish men (age 21). In contrast with the other groups, poorer relationships with classmates were associated with an increased likelihood of HED (age 30) for Swedish women only. For women, the combined effect of having both daily peer contact and living in Finland for HED at age 42/43 was statistically distinguishable from a pure additive effect. Conclusions: Micro and to a lesser extent macro level contexts are associated with heavy episodic drinking well into adulthood. The most relevant processes in the adolescent microsystem occur in family and peer settings. However, long-lasting protective or risk-raising effects between different settings and later HED were not found. Promoting good relationships across different contexts during adolescence may reduce the incidence of HED in adulthood.

Although centred on torture and resource wars, Fred Khumalo's 2015 short story "Water No Get Enemy" begins with bad food. The unnamed narrator establishes a frame setting by joshing about the only recommending feature of the Ekhaya Jazz Restaurant in Johannesburg: "consistency. Take their steaks. [... A]lways huge and rubbery, and unfailingly, excessively salty."1 If Joseph Conrad's frame narrative in Heart of Darkness situates the ship as a key space of imperialism, Khumalo's mocking opening lines take the restaurant as a paradigmatic space of neoliberal-era culture. The frame narrator likewise describes Ekhaya's neighbourhood, Yeoville, in culinary terms: Next door is Kin Malebo Restaurant, where you can see the Congolese in colourful clothes, doing their kwasa-kwasa. [... A]cross the road, [...] Nigerian brothers [are] arguing at... the entrance to the Rotisserie. [...] Zimbabweans are quaffing quart after quart of beer at the Londoner, or Times Square next to it. On the pavements, chaps from Mozambique are roasting chicken gizzards. [...] All these things combined are a boon to the armchair traveller who, without having left South Africa, can have a peek into the lives of "others."2 Khumalo posits that globalisation in Johannesburg can be understood via food. Euroamerican gastronomic pretension appears, but under erasure: the "Londoner" and "Times Square" are characterised as sites of Zimbabwean immigrant presences. These pubs join a pan-African spread of casual eateries, street food, and people. Even as Khumalo's narrator celebrates such African cosmopolitanism, they lament that "Ekhaya is the only remaining South Africanowned business in the legendary Rockey Street strip," implicitly blaming 1 Fred Khumalo, "Water No Get Enemy," in Water: New Short Fiction from Africa, ed. Nick Mulgrew and Karina Szczurek (Oxford: New Internationalist, 2015): 72. 2 "Water No Get Enemy," 72. immigrants from elsewhere in Africa.3 The gastronomic gaze evidences both the positive energies and the fissures of a globalised Johannesburg. "Water No Get Enemy" goes on to situate food as a flashpoint within resource conflicts. As Ekhaya regular Guz-Magesh remarks, "We've had the land wars, the gold wars, the diamond wars, the oil wars [...] What do you think the next war is going to be about? [... I]t will be about food, but there can't be no food without water. The next war will be about water."4 Guz-Magesh frames resource struggle as the driving force of history, linking territorial colonialism to mineral extraction, food politics, imperial penetration of and control over nature, and finally water. History repeats with a difference: the colonial imperative to extract resources persists into the neoliberal era in forms such as global agribusiness, registering a longer series of processes continuous with the recent and specific terror of climate change.5 Fixated on food and water, "Water No Get Enemy" offers attention to consumption and extractivism as ways to consider neoliberal-era relations of cultural dynamism, violence, and material inequality. Thinking from the restaurant, Khumalo's story exemplifies how food connects personal and embodied experiences-hunger, appetite, gustationto planet-scaled dynamics such as globalisation, migration, and environmental crises. Following a similar logic, this special issue takes food, eating, and their meaning-making as points of entry to bring together interlocutors from across the humanities in particular, as well as the qualitative social sciences and creative arts. These producers of knowledge trace intersections among identity, semiotics, culture, political and economic struggles, and consumption as expressed through food cultures in contemporary South Africa. The issue probes how a focus on food might open up new lines of inquiry and fresh inter- Adding to this reconceptualization of nonhuman/human relationships, 'food studies' is an increasingly recognised analytic in the humanities and social sciences, motivating new approaches to interdisciplinary inquiry. Such work has a long history traceable to figures such as the anthropologist Claude Levi Strauss, the philosopher and cultural studies scholar Roland Barthes, the economist Amartya Sen, and the anthropologist Arjun Appadurai. The consolidation of a field called 'food studies' in North America was influenced by thinkers such as Sidney Mintz, whose Sweetness and Power (Viking Penguin, 1985) assayed the now-popular method of tracing globalisation and empire through a single commodity (sugar); Warren Belasco, whose Food: The Key Concepts (Berg, 2008) situated the study of food in relation to culture and environment; and Sherrie Inness, editor of Kitchen Culture in America (University of Pennsylvania Press, 2001), which laid bare the role of food advertising in gender ideologies. North American food studies has been pushed to engage further with race, gender, sexuality, and transnationalism by scholars such as Anita Mannur, Kyla Wazana Tompkins, Psyche Williams-Forson, Sarah D. Wald, and Allison Carruth. Recent food studies work includes dedicated journals, special issues, anthologies, conferences, research and teaching sites, podcast series, and digital media. Scholarship has been supplemented by accessible non-fiction by authors including Michael Pollan, Jonathan Safran Foer, Barbara Kingsolver, and Michael Moss, as well as numerous documentary films. Many edited volumes and special issues published in the Global North, however, focus on areas within the Northeven as they offer useful theoretical and methodological leads for comparative work on food in relation to race, gender, subaltern agencies, and dias-pora.8 Other collections do engage Global South spaces directly, an encouraging direction for further growth in food studies.9 Scholarly monographs, too, have pushed food studies to unpack questions of globalisation and geopolitical power, and to recenter Global South spaces and stories.10 Southern Africa, however, has received little attention in the burgeoning food studies scholarship in the American academy (notwithstanding important counterexamples by scholars such as Evan Maina Mwangi and Jonathan Bishop Highfield). Within South Africa itself, there is a relative dearth of research on how food, eating, and cooking provide signifying sources for identity-construction, or on the social and cultural meanings associated with foodways and cuisines. Even as food features prominently in southern African fiction-writing and material culture, only a few humanistic or social-scientific journal issues in recent years have focused on South African food/human relationships.11 There is a significant volume of research on food in South Africa as elsewhere in the Global South; however, its disciplinary and epistemological influences are often the natural sciences, positivist paradigms, development studies, 'Food sovereignty' and 'food justice' paradigms, on the other hand, critique neoliberal food security approaches. Drawing on Marxian political economy, such approaches highlight the agencies of workers and poor people rather than their passive or supplicant status. Because colonialism and capitalism have operated by destroying the webs of life that Indigenous and marginalised populations rely upon for sustenance, the idea of rebuilding 'food sovereignty' has become crucial to environmental justice writing and activism in places as different as South Africa and Native North America. The South African Food Sovereignty Campaign (safsc) emerged in 2015 as a collaboration among over 50 existing groups representing environmental justice, the solidarity economy movement, waste pickers, landless peoples, and more. For the safsc, 'food sovereignty' means not just "technical solutions for households to access food" but a deeper "transformation of our food system by breaking the control of food corporations, repositioning the state to realise the Constitutional right to food and [...] the emergence of [...] alternatives from below," a united means to address intersecting factors of corporate power, climate crisis, and injustices in land distribution.12 The emergence of this umbrella organisation, uniting activist groups in a common pursuit, points to how a focus on food allows thinking at the nexus of capitalism, culture, colonisation, and environmental change. Paradigms such as 'food sovereignty' and 'food justice' have anchored analyses of food inequalities in a corporate capitalist system that has become increasingly exploitative. This justice-based approach to the reasons for food shortages, poor nutrition, and starvation is often the explicit or implicit foundation of 'critical food studies,' a growing area within South Africanist humanities and social sciences. However, rather than dealing simply with people's food needs for survival, health, productivity, and social justice, a humanities-inflected approach focuses on past and present complexities of human relationships to food. As contributions to this special issue demonstrate, this involves registering and exploring how cultural and social meanings attach to food items, cuisines, and modes of food production. Even the current hegemony of the corporate food system (with its norms of efficiency and rapidity) can-through a critical social science and humanities lens-be revealed as historically and culturally shaped. Being able to eat quickly prepared and highly processed food, food so packed with preservatives that it lasts for months, or exorbitantly priced food promoted as high cuisine, are not markers of an abstract 'progress' or development. They are symptoms of a specific economic system, and the worldview and tastes that both shape and are shaped by this system. Such social and cultural norms are what the humanities and critical social sciences explore. By tracing issues such as how we respond to certain foods emotionally, culturally, psychologically, and sensorily; what we regard as 'good' food; and how belonging and group identification form around eating and buying certain food; humanities-inflected work can contribute to identifying and explaining human consciousness, social processes, and cultural dynamics. --- Issue Contents Building on and moving beyond prior work in food studies, this issue asks what a 'South African Food Studies' looks like by offering scholarly articles and creative pieces engaged with food, culture, and power in southern Africa from a variety of disciplinary, interdisciplinary, and antidisciplinary perspectives. Attentive to questions of knowledge sovereignty, the issue spotlights contributors who are from and/or based in South Africa, and features work by emerging scholars alongside established academics. In addition to exploring how global dynamics such as neoliberalism are expressed in food culture, the issue demonstrates the utility of a 'food studies' lens for exploring community, identity, politics, and power across a range of cultural formations in southern Africaincluding Indian South African identities informed by histories of indenture, Somali and multi-ethnic immigrant community formation, Black middle-class experiences, queer perspectives on organic farming, and the gender dynamics of breastfeeding debates. To bridge gaps between academic and public discourse and to exemplify the wide range of methods enabled by a food studies approach, the issue features several contributions that take creative, personal, and/or anecdotal approaches to food studies, integrating these methods with more traditional research. The first two pieces, by some of the region's leading feminist academics, demonstrate the centrality of feminist politics, theorising, and research to critical food studies. Both trace pivotal lines of inquiry in this transdisciplinary area, such as: How can research into food deepen explorations of what makes us human? Which political, ethical, and epistemological issues underlie decisions or obligations about what to eat? How have food production and cooking been at the centre of gendered, racialised, colonial, and anthropocentric struggles? What roles have Black women especially played in securing food for communities and families, and how has the gendered exploitation of their labour and energies been connected to their struggles against industrial agriculture, corporate food, and the complicity of local governments and the state with extractivist global capitalism? To introduce such questions, the issue begins with a short essay by the late Elaine Salo (1962Salo ( -2016)), an interdisciplinary feminist whose work focused on women's economic, political, and reproductive agencies in the Western Cape, South Africa, and Africa at large. Confronting the neglected roles of African women in food production and struggles around food (rather than seeing women as pawns in patriarchal and class systems), Salo affirms an African feminist approach for recognising women's struggles in intersecting systems of power. An early version of Salo's essay was first published on the Critical Food Studies Programme website; the publication of the revised essay in this issue recognises her pivotal role in South African scholarship on food. Like her other work on women's struggles and agencies, Salo's "Food is an African Feminist Issue" raises provocations for furthering feminist food studies. One key provocation concerns how researchers and writers can sensitively acknowledge the contextually distinct power relations in which women are situated in foodways and discourses around food. For example, by underscoring African women's centrality in struggles against industrial agriculture, Salo draws attention to their agency as ecofeminists within an academic context in which ecofeminism is often seen as an initiative of the North. Similar concerns animate Yvette Abrahams' personal essay "Becoming Food: Queer Indigenes Saving the Planet," which testifies to her years of work as an activist, intellectual, ecofeminist, and organic farmer. Abrahams operates from the premise that just as important as "eating well" is "becoming food," recognising the human body as future compost-and recognising the importance of the often-neglected soil and its tenders, the earthworms. Thinking with the work of queer farmer Eve Balfour, often regarded as the mother of organic farming, Abrahams argues for recognizing earthworms as queer figures who maintain soil health and thus make life possible. In compellingly commonsensical ways, Abrahams also registers the philosophical depth of Indigenous and "posthumanist" explorations of decolonial human/nonhuman relationships through the lenses of eating, cultivating, and decomposing. Picking up threads raised by both Abrahams and Salo, Donna Andrews and Lauren Paremoer take up the challenges of seeing "food as an African feminist issue" in their courageous intervention into 'politically correct' discourses around women and food sovereignty. "Becoming food" takes on an entirely different meaning as Andrews and Paremoer examine recent debates around breastfeeding with an eye to how advocacy for exclusive breastfeeding can constrain the agency of biological mothers. Compared to other African countries, South Africa has a very low breastfeeding rate. Drawing on food sovereignty paradigms, many progressive scholars, policy researchers, activists, and media accounts have responded to this in two ways: infant formula brands associated with global corporations such as Nestlé have been exposed for pursuing profits at the expense of infant nutrition, and scholarship and health media have affirmed the nutritional value of breastfeeding. In the face of these compelling pro-breastfeeding discourses, Andrews and Paremoer's co-authored article, "'Breast is Best'-Really? Thinking through the Intersection of Food, Agency, and Hegemony" exposes how seemingly progressive arguments can become hegemonic. The authors demonstrate how pro-breastfeeding positions can erode women's agency through three case studies: campaigns against manufacturers of baby formula in the 1970s, early 2000s advocacy for hiv-positive mothers to breastfeed, and recent discourses on breastfeeding as the 'first act of food sovereignty.' Highlighting gender expectations and contextuallycontingent meanings of milk from the breast, Andrews and Paremoer illustrate how such campaigns normalise the subordination of certain gendered bodies in the name of public health, the best interests of the child, and/or social res-Matatu 54 (2023) 1-15 istance to Big Food. Andrews and Paremoer also disrupt the tendency of food studies to idealise movements such as food sovereignty, pointing to potential fractures and complications within food sovereignty perspectives. Food sovereignty platforms that claim to defend human rights may, the authors suggest, also undermine certain human beings' rights to choice and agency when people are instrumentalised as 'good mothers' or food-corporation-resisting bodies. The next two essays each unlock complex meanings of food in southern Africa by using qualitative case studies of food-related public acts to scrutinise factors such as immigration, transnationalism, race, class, and gender. In "Black South African Middle-Class Engagements with Social Media and Digitised Food Cultures," Pamella Gysman uses a sociological approach to examine the construction of identities through food-related social media. Applying Pierre Bourdieu's theory of class distinction, Gysman focuses on two aspects of Black middle-class habitus as performed on Instagram: "eating without food," or self-presentation in food-related spaces with actual food absent; and "the cultured palate," or the presentation of knowledge based on culinary experiences, travel, and other forms of cultural capital. Examining how social media influencers utilise these paradigms, Gysman demonstrates that performative acts related to eating enable the creation of Black middle-class identities beyond stereotypes that associate the Black middle class with overconsumption. She therefore expands a key argument of Williams-Forson's study of Black people and food in the United States, insisting that "you cannot forget that on Black people's bodies are written our food legacies-centuries of trauma and desecration, but also centuries of triumph and agency."13 Anyona Ondigi's article "The Languages of Food and Imagined Communities at a Somali Restaurant in Bellville, South Africa" uses an ethnographic approach to explore food culture in one of South Africa's largest hubs of migrant settlement and entrepreneurship. Pursuing an immigrant eating space (not unlike those depicted by Fred Khumalo with which we began), Ondigi shows how the many small ethnic restaurants burgeoning throughout South Africa might be seen as vibrant sites of transnational (not merely ethnic or national) communication, re-identification, and belonging. Despite the xenophobic responses of many South Africans to migrants from other African countries, and irrespective of many migrants' consolidation of in-group solidarity in reaction, the 'ethnic restaurant' can nonetheless function as a powerful and influential site of collective eating, sharing, socialisation, and conversation-Matatu 54 (2023) 1-15 both within and beyond particular ethnic, national, and religious groups. This analysis recalls Jürgen Habermas' notion of a public sphere, highlighting how certain food events and sites bring individuals together to debate social problems and, through that, influence public opinion and effect cultural and political changes. The next two pieces integrate scholarly and creative or personal approaches, broadening the critical food studies conversation beyond academic discourse. Supplementing Ondigi's attention to communities of recent immigrants from Somalia and elsewhere on the African continent, Pralini Naidoo turns to the long and complex histories of South Asian arrivals in South Africa. Her essay "The Unwritten: Exploring Food Stories Beyond and Between the Lines" traces how legacies of indenture and immigration inform contemporary Indian South African food cultures. Contemplating the relationship between written and unwritten modes of transferring culinary knowledge across generations, Naidoo teases out complexities and historical hierarchies behind the multiple identities and food cultures often homogenised as 'Indian' in contemporary South Africa. She takes personal anecdotes, family history, recipes, and cookbooks all as objects of analysis-exemplifying the expanded archives of literature and culture made possible by a 'food studies' approach. In acknowledging cultural knowledge that has been lost, amalgamated, or transformed as well as that which can be traced, Naidoo affirms culinary heritages and identities as dynamic-drawing on layers of entangled histories-rather than static or purist. She asserts the importance of food for constructing a sense of home in the wake of such entangled histories. Following Naidoo's personal/historical essay, Leonie Joubert takes a combined creative, journalistic, and academic approach to exploring how alimentation operates as a form of power on both global and embodied scales. Joubert's protest artwork "Hit Me(n): Profiting from the Corporate Capture of the Dopamine Drive since 1492" combines seven original drawings and narrative text to critique what Joubert calls the "hunger-obesity poverty-paradox" through a focus on sugar. Linking the negative health impacts of sugary massproduced foods to histories of how enslaved labour enabled mass sugarcane cultivation and sped globalisation, Joubert makes a case for reframing problems such as obesity in terms of corporate rather than individual culpability in the accompanying article "Hit Me(n): Protest Art as a Communications Tool to Reframe the Hunger-Obesity Poverty-Paradox in the Global South." Joubert's drawings, meanwhile, work through a provocative logic of substitution to ask readers to consider how the harmful effects of sugary foods might be reconsidered in comparison to other 'unhealthy' inputs-whether cocaine, cigarettes, spilt petroleum, or Facebook. While accompanied by an academic Matatu 54 (2023) 1-15 rationale, Joubert's protest art helps to expand the issue's purview and language beyond the academic space, reminding us of the importance of inviting broader publics to consider how food cultures in South Africa intertwine with the forces of globalisation. The final piece in the special issue addresses how the power dynamics of food access in South Africa-explored throughout the issue-have been amplified by the covid-19 pandemic. Food access for many South Africans was imperilled during the early phases of the pandemic, as Rejoice Mazvirevesa Chipuriro and Kezia Batisai discuss in "House of Hunger: The Weaponisation and Politicisation of Food (Protests) in South Africa During covid-19." Protests highlighting hunger, moreover, were criminalised as "riots." Arguing against this dismissive treatment, Chipuriro and Batisai illustrate how food protests both reveal underlying structural problems in the food system and exemplify a "hyper-visualisation" of Black bodies as unruly. They therefore confront tropes and misconceptions explored by African American scholars such as Williams-Forson, and also highlight the importance of critically analysing economic, political, and discursive usages of subaltern subjects' bodies. The problematic construction of the Black body, the authors argue, threatens to block pathways of change in the food system, and in other relational worlds where certain subjects are othered in order for the privileged to consume food and other resources far in excess of what human beings need. Chipuriro and Batisai's essay concludes the special issue with an emphasis on how the power relationships indexed through narratives about food remain more important than ever in our precarious present. --- Conclusion Economist Amartya Sen recognised decades ago that hunger is never simply experienced as a bodily sensation of want, but also as a psychosocial experience of shame and low self-esteem.14 Human relations to food are never purely pragmatic or biological, but also deeply social and cultural. Humanities work on food therefore expands our understanding of a range of experiences, historical processes, and social norms. As a central facet of our lives, which all humans both think about and engage with daily, food is intricately enmeshed with our desires and tastes; our relationships with others and our statuses in 14 See Amartya Sen, Poverty and Famines: An Essay on Entitlement and Deprivation (New York: Oxford up, 1992). the world; our senses of self and others; our fantasies, memories, and ways of reasoning; and our perceptions of bodies, beauty, and health. By using food as a lens in our approach to society, history, and culture, we are able to extend and deepen our analysis as humanities and critical social science thinkers-in the South African context and beyond. Even as this issue centres South Africa, the contributed pieces also draw attention to historical and current flows across the southern African region and globally. Indeed, food studies approaches offer powerful means to connect particularised concerns about personal identity, the body, and the kitchen with broader national and transnational regimes of power and resistance-such as agribusiness corporations, food supply chains, global patterns of environmental degradation, and the covid-19 pandemic. Also important to this work is foregrounding the food cultures of socially marginal and radical groups. In sum, this special issue leverages a focus on food cultures in South Africa to open up fresh approaches to the most urgent and challenging issues of our times. --- Works Cited

Introducing a special issue of Matatu titled "South African Food Studies," this essay argues for the importance of food as a lens for understanding contemporary culture and society. More specifically, the essay advocates for recentring Global South contexts-in this case South Africa-in a 'food studies' conversation that has often been dominated by the American academy; it also underscores the vitality of the humanities, qualitative social sciences, and creative arts for transcending reductive 'food security' paradigms often applied in the Global South. The essay first examines the short story "Water No Get Enemy" by South African writer Fred Khumalo, introducing how a focus on food and eating can illuminate globalisation, xenophobia, resource conflict, and environmental change. From here, the authors introduce the evolving field of 'food studies,' then outline the eight academic, personal, and creative pieces that constitute this special issue, all authored by contributors from the African continent. Issues raised include the gendered and queer politics of food, breastmilk, and soil; the ongoing coloniality of neoliberal approaches to food inequality; the burdening of Black bodies; the role of so-called 'ethnic restaurants' in building transnational Downloaded from Brill.com 05/10/2024 02:20:04AM via free access stanley, lewis and mafofo Matatu 54 (2023) 1-15 and multi-ethnic communities; and the heightened stakes of food access during the covid-19 pandemic.

IntroductIon Infectious diseases are responsible for approximately half of child deaths worldwide, 1 where pneumonia and diarrhoea are two of the leading causes. 2 3 Apart from deaths, several diarrhoeal episodes lead to long-term nutritional deficits, whereas childhood respiratory infections cause increased risk of reduced lung capacity. 4 5 The greatest proportions of severe episodes of these infections are experienced in the south Asian and African regions and within these regions, the prevalence of these diseases is even higher for poorer countries and among disadvantaged children. 6 7 Considering the consequences of diarrhoea and acute respiratory infection (ARI) episodes on mortality and long-term adverse health outcomes, numerous studies have been conducted to identify associated risk factors. Existing literature suggests that younger children are more likely to experience infectious diseases than older cohorts, and consequently, a higher proportion of diarrhoea and pneumonia-related deaths happen in the first 2 years of life. [7][8][9] Due to differentials in food intake, standard of living and availability of healthcare facilities, urban-rural variations Open Access are also evident in childhood morbidity. 10 11 Educated mothers are knowledgeable about healthy environments and possible risks of their children being exposed to infectious diseases, and hence, occurrence of infectious diseases is significantly associated with the educational status of mother. [12][13][14] In addition to formal education, access to mass media helps to create awareness around communicable diseases. [15][16][17] Furthermore, poorer individuals within impoverished settings face a relatively higher burden of infectious diseases, and children in wealthy families are more protected from diarrhoea and ARI. [18][19][20] At the end of the Millennium Development Goals (MDG) in 2015, the development policies emphasised reducing the percentage of people living in households lacking a sustainable environment in terms of durable housing structure, sufficient living area, access to safe water and access to improved sanitation. These policies were developed considering the fact that a large proportion of diarrhoea-related mortality is attributable to either unsafe drinking water, inadequate sanitation or insufficient hygiene. 21 Investigating 171 Demographic and Health Surveys (DHS) in 70 low-income and middle-income countries over the period 1986-2007, researchers 22 concluded that access to improved sanitation and water was associated with lower risk of childhood diarrhoea. Similar relationships have been observed elsewhere. [23][24][25] As a part of Goal 3 of the Sustainable Development Goals (SDG), the UN is aiming to end epidemics of water-borne and other communicable diseases by 2030. Goal 6 emphasises ensuring the availability of water and sanitation for all. 26 A review of published articles 27 concluded that residential crowding significantly increased the risk of severe respiratory disease. Using solid fuel for cooking is a major source of household air pollution and responsible for a variety of respiratory diseases. [28][29][30] The main aim of this study was to advise on the relative importance of the SDG, in the light of evidence around the prevalence of diarrhoea and ARI among preschoolers from three south Asian countries-Bangladesh, Nepal and Pakistan. The first objective was to compare the prevalence of diarrhoea and ARI across the three countries. The second objective was to identify characteristics of children with relatively higher risks of diarrhoea and ARI. The third was to compare the association of potential predictors with the prevalence of diarrhoea and ARI. Of particular interest in the light of the SDG is the relative importance of sustainable household environment, in terms of safe drinking water and toilet facilities, compared with other predictors such as maternal education and household wealth. A second aim was to provide suggestions regarding the development of feasible and effective plans that fit with the SDG and are likely to reduce childhood diarrhoea and ARI in the region. --- data and methods This study achieved the objectives by using the latest releases of nationally representative and cross-sectional DHS datasets from three south Asian countries-Bangladesh (2014), Nepal (2011) and Pakistan (2012). These developing countries are home to 400 million people (Bangladesh 169 million, Nepal 32 million and Pakistan 199 million) 31 and share common historical, social and cultural background. In terms of per capita GDP, they trail the world (Pakistan 152nd, Bangladesh 156th and Nepal 172nd). 32 Significant variations also exist among the countries in terms of child morbidity outcomes as well as a range of exposures to poverty and unhygienic household environment. The DHS consists of a nationally representative sample of households obtained through a two-stage stratified sampling procedure. First, sample sizes in terms of number of households (for Bangladesh 18 000 households, for Nepal 10 826 households, for Pakistan 14 000 households) were calculated to provide reasonable precision for the survey indicators. In the first stage, each of the countries were divided into strata and a sample of enumeration areas (EAs) were selected independently from each stratum using probability proportional to size. The EAs were considered as the Primary Sampling Units. In the second stage, a systematic sampling technique was employed to select a fixed number of households from each of these EAs. Finally, ever-married women aged between 12 and 49 years living in the selected households were approached for interview to collect necessary information. The DHS enjoy high response rates and provide cleaned data for secondary analysis. This study excluded children from women who were not de-jure resident at the time of interview. Episodes of diarrhoea and ARI in 23 940 children (7760 from Bangladesh, 5140 from Nepal and 11 040 from Pakistan) under the age of 5 years and born to the selected women were analysed. To ensure the representativeness of the sample at various levels, sampling weights are included in the data for each sampling stage and cluster based on sampling probabilities. The sample weights were incorporated into the current analyses. The surveys have attractive features that make them appealing for quantitative analysis: they collect information about the morbidity status of children and cover a wide range of variables regarding the children, their parents and the households they live in. The variables used in this study will be described in the next section. --- dependent and independent variables The dependent variables in the study come from mother's responses to questions on recent (within 2 weeks preceding the survey) episodes of diarrhoea and ARI of their children aged below 5 years. The DHS identify a child experiencing ARI by asking the mother whether the child has been ill in the 2 weeks preceding the survey with a cough accompanied by short, rapid breathing or by difficulty in breathing that the mother considered to be chest related. 14 The children were categorised as those suffering or not experiencing these episodes within 2 weeks before the survey. Similarly, the DHS identify a child experiencing diarrhoea by asking the mother whether the child had diarrhoea in the last 2 weeks preceding the survey. Open Access The set of independent variables considered as potential predictors were decided on using the existing literature, availability of information in the survey datasets and the knowledge of the researchers of the study region. To compare the prevalence of diarrhoea and ARI across age groups, children were categorised as those aged less than 1 year, between 1 and 3 years and from 3 to 5 years. A similar categorisation is used by other researchers. 17 Households were also categorised as being in either an urban or rural setting. Educational attainment and access to mass media have been considered as a proxy of knowledge and understanding of mother regarding exposure to infectious diseases and their consequences on their children. 15 16 Mothers were categorised as those with no or incomplete primary level, complete primary to incomplete secondary level and complete secondary or above level of education. Mothers were also classified on the basis of having or not having access to any of radio, television or newspapers. Significant positive associations between household economic status and prevalence of infectious diseases were observed in previous studies. 14 20 As a proxy of household economic status, this study used the wealth score created by the DHS and calculated using principal components analysis from variables comprising household construction materials (roof, ceiling and floor), possessions (televisions and bicycles) and dwelling characteristics (source of drinking water and sanitation facilities). Details about the calculation of the wealth scores is available. 33 On the basis of wealth score, children were classified as those coming from a household classified as low (lower 40%), middle (middle 40%) and high (upper 20%) wealth categories. Source of drinking water, type of toilet facility, crowding, type of cooking fuel and type of floor material were considered as indicators of sustainable household environment. Pipe, borehole, protected dug well, spring or rainwater are considered as improved sources of drinking water. 33 Households were categorised as those having or not having an improved water source located on the premises. Sanitation using technologies such as flush toilet, ventilated pit latrine, traditional pit latrine with a slab or composting toilet were considered as improved. Households were categorised as those with ideal (improved unshared), moderate (unimproved unshared or improved shared) or worst (unimproved shared or no facility) toilet facilities. The number of adults per living room indicates crowding, and households were categorised as those with up to two adults per living room and two or more adults per living room. On the basis of existence of cooking practice, the households were categorised as ideal (cooking outside the house or not using solid fuel while cooking inside) or not ideal (using solid fuel while cooking inside the house). Finally, houses were categorised as having the floor made of mud or else. All information was extracted from the datasets mentioned in the previous section. --- statistical analysis The dependent variables (prevalence of diarrhoea and ARI), as well as the predictor variables considered in the study, are categorical. Bivariate <unk> 2 analyses were carried out to compare the prevalence of the diseases among the levels of the selected predictors. Given the dichotomous nature of the dependent variables, multiple binary logistic regression models were fitted to assess the association between selected predictors on childhood morbidity. The model considers logarithm of odds (ratio of the probability of occurring to not occurring) as a linear additive function of the predictors. Exponentials of the estimated parameters referred as the OR estimate the changes in the odds with unit change in the predictors (for continuous predictor) or changes in the level of predictors compared with baseline (for categorical predictor). 34 Separate multiple logistic regression models were fitted to the data from individual countries. To focus the models for each disease, the model for diarrhoea excluded the variable cooking fuel as predictor, whereas the model for ARI excluded source of drinking water and type of toilet facility as predictors. Statistical analysis was conducted in SPSS 21.0 35 using a weighted analysis to account for the survey weights and clustered structure of the sample. --- results Prevalence of diarrhoea was considerably lower for the children in Bangladesh (5.7%) with respect to Nepal (13.8%) and Pakistan (22.5%). A considerably higher percentage of children from Pakistan were experiencing ARI (15.9%) than Bangladesh (5.4%) and Nepal (4.6%). Percentages of children experiencing comorbidity (defined in this study as presence of both diarrhoea and ARI) were highest in Pakistan and lowest in Bangladesh (figure 1). Distributions of respondents for various levels of selected predictors are presented in table 1. Age distributions of children were consistent over the studied Open Access countries. Most of the children in the studied settings lived in rural areas, ranging from 70.3% in Pakistan to 90.6% in Nepal. The percentage of mothers with both lower (no or incomplete primary) and higher (completed secondary or higher) educational levels were highest in Pakistan (62.5% and 17.9%), whereas the percentage of mothers with no or incomplete primary level of education was lowest in Bangladesh (32.7%). The highest proportion of children whose mother had access to any mass media was from Nepal (84.4%). Among the studied countries, percentages of children from households using improved and unshared toilet facilities were lowest in Nepal (35.9%). Using solid fuel while cooking inside home was rare in Bangladesh (9.6%) and common in Nepal (57.4%) and Pakistan (59.8%). Among the studied settings, the percentage of children from household sharing more than two adults a room was highest (71.7%) in Pakistan. In the bivariate analyses, children experiencing episodes of ARI (within 2 weeks before the survey) were more likely to experience diarrhoea and vice-versa (table 2). Older children were significantly less likely to experience Open Access either episode than the younger. With the exception of episodes of ARI in Nepalese children, educational status of mother showed a significant association with the prevalence of the episodes of both infections. A mother's access to mass media showed significant association with both episodes in Bangladesh only. Except for the prevalence of diarrhoea in Bangladesh, the wealth status of a household showed significant association with childhood diarrhoea and ARI. Significant associations between place of residence and prevalence of ARI were also observed for children in Bangladesh and Pakistan (p<unk>0.05). In all except one setting, the source of drinking water did not show any significant association with diarrhoea or ARI, whereas type of toilet facility showed a significant association with the prevalence of diarrhoea in Nepal and Pakistan. The number of adults per living room showed a significant association with the prevalence of diarrhoea in Bangladesh and Nepal. Children from households using solid cooking fuel inside the house were significantly more likely to experience ARI with respect to those from households not using solid fuel for cooking inside the home. Adjusted ORs for experiencing diarrhoea and their associated CI for the predictor variables are presented in table 3. Once the effect of other variables was controlled for, in all the three countries, age of children and educational status of mother showed significant association with prevalence of diarrhoea. In Bangladesh, Nepal and Pakistan, children below 1 year of age were 63% (OR 1.63, CI 1.23 to 2.16), 219% (OR 3.19, CI 2.51 to 4.05) and 156% (OR 2.56, CI 2.26 to 2.91) more likely to experience diarrhoea with respect to those aged above 3 years (the reference category). With respect to the reference category, children aged between 1 and 3 years were also significantly more likely to experience diarrhoea. In Bangladesh, children from mothers with no or incomplete primary and incomplete secondary educational levels were 70% (OR 1.70, CI 1.16 to 2.49) and 69% (OR 1.69, CI 1.20 to 2.38) more likely to experience diarrhoea with respect to those from mothers with secondary or higher level of education. In Pakistan, children from households categorised as low and middle wealth status were significantly more likely to experience diarrhoea with respect to those with high economic status. However, for the children from Bangladesh and Pakistan, the wealth status of household was not significantly associated with the prevalence of diarrhoea (p>0.05). Children from households using worst toilet facilities (in terms of improvement and sharing status) were more likely to experience diarrhoea; however, the association is statistically significant for the children in Nepal only. Adjusted ORs for experiencing ARI with associated CI for the studied variables were presented in table 4. Like diarrhoea, younger children were significantly more likely to experience ARI than older children. In Bangladesh and Pakistan, children of mothers with incomplete primary or incomplete secondary level of education were more likely to experience ARI than those from secondary or higher level of education. In Pakistan, with respect to the children of mothers with secondary or higher level of education, those of mothers with incomplete primary and incomplete secondary level of education were 35% (OR 1.35, CI 1.13 to 1.62) and 24% (OR 1.24, CI 1.04 to 1.49) more likely to experience ARI. However, the relationship of maternal education and ARI of children is not significant in Nepal. In Bangladesh and Nepal, children from households categorised as low or middle wealth status were significantly more likely to experience ARI with respect to those with high status. However, the association of wealth with ARI was not significant for Nepal. In Bangladesh, children from households not using ideal cooking facilities (using solid fuel while cooking inside house) were 62% (OR 1.62, CI 1.22 to 2.16) more likely to experience ARI; however, the association was not statistically significant in Nepal and Pakistan (p>0.05). Over the countries, neither the degree of crowding nor the household floor material showed consistent association with either diarrhoea or ARI. --- Open Access --- dIscussIon Prevalence of two major infectious diseases, diarrhoea and ARI, is relatively high among young children in the South Asian region. The study was conceived with objectives of comparing the association of potential predictors, in particular sustainable household environment and maternal education, on disease episodes among preschoolers from three South Asian countries, Bangladesh, Nepal and Pakistan. The highest prevalence rates were in Pakistan for both diarrhoea and ARI, and in Nepal, for diarrhoea. Bangladesh is considered as a paradox in terms of good health outcomes despite economic poverty. 36 37 Even the disadvantaged children from Bangladesh (from households categorised as poor and using unimproved water and toilet facilities) possess lower risk of experiencing diarrhoea with respect to economically advantaged children in Nepal and Pakistan. Similar outcomes have been reported, 38 where it is mentioned that Pakistan lags behind the MDG in many aspects including child health. Though the prevalence of ARI is low in Nepal, prevalence of diarrhoea is higher in the country. This may have resulted from lack of sustainable household Open Access environment (defined as improved water source and hygienic sanitation) or knowledge of healthy practice of mother (measured by level of schooling). Rural-urban setting was not found to have a significant effect on the prevalence of either diarrhoea or ARI. Source of drinking water, number of adults sharing a living room and floor materials all showed inconsistent effects on the prevalence of childhood diarrhoea across the countries studied. In all the three studied countries, bivariate analyses showed that the presence of diarrhoea significantly increases the likelihood of the prevalence of ARI and viceversa. However, it is important not to include diarrhoea in the multiple model of ARI prevalence and vice versa, because there is a high degree of overlap in risk factors for diarrhoea and ARI, for example, age of child, crowding, poor housing. Furthermore, the short recall period employed in the DHS (2 weeks) means that the data are clearly cross-sectional and the lag effect of one disease on another cannot be measured with the data at hand. Similar results were observed in previous studies where the epidemiology of diarrhoea and ARI overlapped. This is highly likely to be due to shared risk factors or compromised immune function. [39][40][41] Health policy needs to take a holistic approach to combatting childhood infections due to the clear presence of comorbidity, at least in the case of diarrhoea and ARI shown in this article. Younger children possess significantly higher risks of experiencing morbidity than older; the result is consistent with previous studies in Bangladesh and elsewhere. 8 14 The immune system may not be developed at earlier ages, and younger children may be infected from unhygienic feeding practices (water, bottles, and so on) and unclean surroundings. Diarrhoea pathogens, like Escherichia coli, are commonly transmitted via impure water, unhygienic utensils or poor food handling. 42 While crawling, children explore their immediate environment and may pick up infections. Inappropriate dietary supplementation may also hinder children's developing immune system, which can be overcome through exclusive breastfeeding for recommended periods of time. Breastfeeding data are available in the DHS; however, it has not been included in the current analyses due to its high degree of confounding with age. --- Open Access Of most interest is the result that maternal educational status showed significant positive influence on reducing the prevalence of diarrhoea and ARI, and that the effects are more evident when the educational attainment is at least secondary or higher. --- data limitations Although the DHS adopted appropriate survey methodology to obtain a representative sample, the final sample does not guarantee complete representativeness at national and regional levels. DHS are cross-sectional surveys which collect information regarding the disease episodes for a single point of time. Such studies can yield information on association but not impact. The seasonal variations in the prevalence of the disease episodes are not addressed in the data. Moreover, the disease episodes are determined on the basis of self-reporting of mothers and not followed by any clinical examination. To reduce the reporting bias due to memory lapse, a short recall period (2 weeks) is considered while collecting morbidity related information. Children who were not de-jure resident at the time of the survey were excluded from the analysis. The exclusion may lead a bias to the outcomes; however, the amount of bias is likely to be small. Finally, the datasets do not represent same time point for the studied countries, though the largest time difference is only 3 years. conclusIon This study focuses on three South Asian countries with relatively higher prevalence of childhood diarrhoea and ARI on the worldwide scale. Sustainable household environment, as mentioned in the SDG, in terms of flooring, water, toilet and cooking facilities was not found to have the greatest association with morbidity. Nor was wealth a main driver, as its association with childhood diarrhoea and ARI was only statistically significant in half of the settings considered in this paper. Of far more importance was the knowledge base of a child's primary caregiver (typically the mother) regarding the potential risks of infection and the impacts of infection on survival and well-being. These results indicate that the SDG of good health and well-being needs to be tackled by incorporating quality health education along with the goals of clean water and sanitation for all and no poverty. According to our findings, maternal education could be effective in reducing child morbidity only when it is at complete secondary level or higher. An effective basic knowledge base formed through policies that incorporate health education in school curricula at primary and secondary level has a good chance of making an impact, especially since the female primary school enrolment rate in Nepal and Bangladesh is already above 90%. Alternative ways to develop knowledge of healthy practices among the mass population could be delivered through mass media. Well-developed motivational programs incorporating mass media, health professionals, community health workers, community leaders, government and non-government organisations may help improve population awareness of the causes and consequences of infectious diseases. These programs may be more essential for the countries like Pakistan where the percentages of children suffering infectious diseases are relatively higher and female school enrolment rate is relatively lower.

Objectives Prevalence of diarrhoea and acute respiratory infection (ARI) is considerably high among South Asian children. The objective of this study is to compare the associations of sustainable household environment and knowledge of healthy practices with episodes of these diseases among the children in the region. Design The study analysed the latest, nationally representative and cross-sectional Demographic and Health Survey data. Setting Data from three countries were analysed: Bangladesh, Nepal and Pakistan. Participants Women aged between 12 and 49 years living in selected households provided information on 23 940 of their children under the age of 5 years. Primary outcomes measures The morbidity status of the children was recorded with respect to episodes of diarrhoea and/or ARI in the 2 weeks preceding data collection. Results Consuming unhygienic drinking water increased the risks of childhood diarrhoea, and use of solid fuel for indoor cooking increased the risk of ARI, across all three countries investigated. However, far more significant were the effects of mother's education, with incomplete primary education leading to an odds of diarrhoea approaching twice that of a mother with secondary education or higher (OR 1.70 in Bangladesh, 95% CI 1.16 to 2.49). Conclusions Results from the current research underline the importance of developing and implementing integrated strategic plans for mothers and children in the countries investigated. Promoting hygienic water and sanitation facilities can help reduce the prevalence of childhood diarrhoea. Replacing indoor solid fuel cooking arrangements with cleaner fuel or more airy conditions can help reduce the prevalence of ARI. However, these strategies need to be integrated with education for women to raise the likelihood that reduced risks are actually realised.

Introduction In the United States, policies in each state increasingly affect the lives of the nation's 22.4 million noncitizens and, in particular, the 11 million who are undocumented (Zong & Batalova, 2016). In 1996, the federal Illegal Immigration Reform and Immigrant Responsibility Act (IIRIRA) and the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) gave states increased discretion over policies that determine the rights and public program eligibility of noncitizens (Bitler & Hoynes, 2011;Fragomen, 1997). While the federal government retains sole authority over immigration law (who can enter the country and their legal status), immigrant laws and policies affecting noncitizens' health care access, education, workplaces, and more are also determined by state-level legislation and regulations (Motomura, 2014;Wallace & Young, 2018). The growing body of knowledge on social determinants of health shows that policies across multiple sectors have an impact on well-being (Apouey, 2013;Marmot & Jessica, 2014;Phelan, Link, & Tehranifar, 2010;Shi, Tsai, & Kao, 2009); the health in all policies approach emphasizes that non-health policies have an impact on population health (Rudolph Linda, Connie Mitchell, Karen Ben-Moshe, & Lianne Dillon, 2013). State-level immigrant policies can extend rights and programs to categories of immigrants otherwise excluded by federal policies, providing them access to critical safety net resources, or they can reinforce federal policies and establish state rules that restrict immigrant rights or program eligibility, further marginalizing low-income immigrant communities. More than any single policy or program, it is a state's combination of policies that shape the context of settlement and incorporation, and reinforce attitudes towards immigrants. Within the different policy contexts across US states, immigrants may be included through extended social and economic rights, such as the ability to pursue higher education, or excluded by lack of protections, such as in the workplace, or excluded through discrimination and active surveillance and enforcement by local law enforcement (Almeida, Katie, Pedraza, Wintner, & Viruell-Fuentes, 2016;Flores, 2010;Kline, 2017). As a result, the determinants of immigrant health vary across states. The relationship between immigrants' legal status and health has increasingly become a topic of public health concern (Castaneda et al., 2015;Hardy et al., 2012;Martinez et al., 2013;Menjivar & Kanstroom, 2014;Torres & Young, 2016;Young & Pebley, 2017). For example, studies have found that undocumented immigrants, compared with citizens and documented immigrants, are less likely to have health insurance and receive timely preventative or, in the case of pregnant women, prenatal care, and are more likely to experience depression or psychological distress (Arbona et al., 2010;Korinek & Smith, 2011;Ortega et al., 2007;Potochnick & Perreira, 2010;Roche, Vaquera, White, & Rivera, 2018;Rodriguez, Bustamante, & Ang, 2009;Stevens, West-Wright, & Tsai, 2010;Vargas Bustamante et al., 2012). While undocumented immigrants are in a position of unique vulnerability (Quesada, Hart, & Bourgois, 2011), other noncitizens may also experience worse outcomes compared with citizens. One study found that Lawful Permanent Residents, who lack full citizenship, experienced worse health in later life than their counterparts who had naturalized (Gubernskaya, Bean, & Hook, 2013). Another found a greater likelihood of depression and psychological distress among all noncitizens, compared with naturalized and US born citizens (Gee, Morey, Walsemann, Ro, & Takeuchi, 2016). In addition, there is evidence that citizens who are in mixed status families, such as US born children with undocumented parents, experience worse outcomes compared with other citizens, such as lower rates of insurance and poor developmental outcomes (Stevens et al., 2010;Yoshikawa, Godfrey, & Rivera, 2008). This growing body of research suggests that the stratification by citizenship and legal status is associated with varying levels of inequality or marginalizationand health disparities -among immigrant populations (Torres et al., 2016). The differences between citizens and noncitizens, and between documented and undocumented noncitizens, vary across place and context. Each legal statusand how it might affect well-beingis shaped by the differential policies that grant rights and opportunities to those who are not citizens (Gee & Ford, 2011;Motomura, 2007) and create inclusive or exclusive immigration climates. In states with more inclusive policies, noncitizens are treated more like citizens, with access to some of the same resources and protections that mitigate discrimination or concerns about enforcement and deportation (Motomura, 2007). In states with restrictive policies, as has been increasingly highlighted in health research, noncitizen groups are more likely to be uninsured and low income, as well as experience greater discrimination and racial profiling due to overall anti-immigrant climates (Anderson & Finch, 2014;Motomura, 2007;Rhodes et al., 2015;Young, Elena, Leon-Perez, Wells, & Wallace, 2017;Young, Elena, León-Pérez, Wells, & Wallace, 2018). While undocumented noncitizens face the most barriers, the policy context also influences the racialization of people of color who are documented or US born and who may experience racial profiling and discrimination despite their citizenship or legal status (Arnold, 2007;Garc<unk>a, 2017;Romero, 2011;Viruell-Fuentes, 2007). Understanding the landscape of state immigrant policies is critical to understanding the factors that shape health among noncitizens and the potential variation in health across US states. To examine the role and influence of state immigrant policies in relation to the social determinants of health, we: 1) provide a framework for understanding immigrant policy as a social determinant of health, 2) apply the framework to measure and assess policies in each US state, and 3) examine state immigrant policies in relation to other state-level social determinants of health. 1.1. A framework of immigrant policy as a social determinant of health State immigrant policies are laws, regulations, and court rulings within various areas of public policy that create differential rights and opportunities based on an immigrant's legal status or citizenship. Immigrant policies play a role at both the institutional and individual levels in the well-being of immigrants (Castaneda et al., 2015;Galeucia & Hirsch, 2016;Torres et al., 2016). They can work by shaping the broad social and economic conditions that immigrants are most exposed to, conditions that are generally recognized as social determinants of health (Marmot & Jessica, 2014). Immigrant policies may also have a more proximal impact in providing access to health care and other health related resources. As shown in Fig. 1, immigrant policy, like other forms of public policy, has direct effects on many of the institutions where immigrants live and work. These institutions shape individuals' access to opportunities that promote health, such as educational advancement, or exposure to circumstances that may harm their health, such as being constrained to a segmented labor market where immigrants with low educational levels are limited to dangerous occupations (Braveman, Egerter, & Williams, 2011;Phelan et al., 2010). The overall set of policies that influence key institutions reflects and helps create an ideological context that can be characterized as a "climate" of settlement and incorporation for immigrants. The resulting climate shapes the racialization of people of color (Garc<unk>a, 2017;Romero, 2011) and creates social and economic security or insecurity, promotes immigrants' certainty or uncertainty about their daily lives, and shapes their perspectives about the trustworthiness of public institutions (Massey & Sánchez, 2010). Social determinants at the institutional and systems level are created by multiple social and economic structures that shape individuals' lives and life chances, from housing markets to workplaces to schools to the criminal justice system. These structures can produce inequitable conditions that result in disparate opportunities based on individuals' legal status, as well as other social categorizations such as race, class, or gender. At the individual-level, distinct and inequitable experiences of social structures also have a social and psychological effect on individuals as they navigate social and economic institutions, develop an identity, and make sense of their position in society (Rachel et al., 2016). This can create general stress and anxiety, as well as influence decisions about health behaviors and public service use. State-level immigrant policies that affect the rights and opportunities of undocumented immigrants currently exist in five broad public policy domains: 1) public health and welfare benefits; 2) higher education; 3) labor and employment policies; 4) drivers' licenses and identification systems; and 5) immigration enforcement. Many areas of policy may have an impact on the health of communities (Rudolph et al., 2013), but it is within these five primary domains that state-level policy makers currently have discretion to create differential access and opportunities based on citizenship or legal status (Motomura, 2014). The policies in these domains function to establish rights, eligibility and other criteria based on a person's citizenship or legal status and include or exclude undocumented and other immigrants through different social and institutional mechanisms. Some state policies explicitly exclude noncitizens by aligning with federal policies that bar individuals from resources or eligibility based on legal status. For example, the federal policy for funding Medicaid excludes coverage for the undocumented and requires a five year wait for Lawful Permanent Residents (LPR). Nineteen states retained these exclusions and offer no medical coverage for low-income pregnant women who are undocumented or recently arrived LPRs. In contrast, inclusive state policies proactively and explicitly opt individuals in, such the 15 states that extend Medicaid to LPR women without the five year wait, despite not receiving federal Medicaid matching funds for their care. An additional 17 states provide Children's Health Insurance Program (CHIP) through a federal option or other coverage to low-income pregnant women regardless of legal status or year of entry (2016). Many state policies are implicitly inclusive or exclusive by establishing requirements that are linked with legal status. For example, states that bar undocumented immigrants from obtaining driver's licenses do so by requiring forms of identification, such as a Social Security number, that the undocumented do not possess. Public health and welfare benefits, access to higher education, and labor and employment protections directly affect access to health care and other social determinants of health (Braveman et al., 2011;Marmot & Jessica, 2014;Rudolph et al., 2013). Driver's licenses/identification and the federal immigration enforcement program, Secure Communities1, represent two areas of immigration policy that have been under debate in recent years in many states that have a more indirect health impact. A driver's license provides an undocumented immigrant with mobility and identification that facilitates access to banking and other economic resources, including a different array of occupational opportunities. In contrast, policies such as Secure Communities required state and local law enforcement to collaborate with federal immigration authorities, creating a "chilling effect" that discourages immigrants from using public services and being in public places (Watson, 2010), in addition to increasing the fear and stress among undocumented immigrants and their families that can have health impacts (Arbona et al., 2010;Cavazos-Rehg, Zayas, & Spitznagel, 2007). While immigrant policies tend to focus on a specific noncitizen group, such as the undocumented or LPRs with less than 5 years of residence in the US, there is evidence that the exclusion or extensions of rights of one noncitizen group has an impact on other noncitizens and individuals in mixed status families. For example, current immigrant policies exist within a broader national context of anti-immigrant climates, creating discrimination against noncitizens in general (Esbenshade & Obzurt, 2008). There is evidence that restrictive policies targeting undocumented immigrants have a "spill-over" effect on documented noncitizens, as well as citizens in mixed status families (Martinez et al., 2013;Rhodes et al., 2015;Stevens et al., 2010;Yoshikawa et al., 2008). This may include undocumented parents being deterred from seeking services for their citizen children due to concerns about enforcement (Rhodes et al., 2015) and immigrants, regardless of legal status, experiencing racial discrimination in a variety of contexts (Anderson & Finch, 2014). In addition, immigrant policy climates place immigrants, and Latinos, in particular, in a position of "racialized legal status" where their race or ethnicity is conflated with being undocumented (Asad & Clair, 2017;Garc<unk>a, 2017). Therefore, while any single policy has the most immediate impact on the group targeted, it also plays a role in shaping the overall context that impacts the experiences of others within a state, placing at social disadvantage both those who are undocumented and documented. Further, immigrant policies across social, economic, and political sectors work together to shape outcomes for immigrants. Much of the recent research on immigrant policy and health has focused on single policies, such as local law enforcement cooperation with national immigration agencies; yet, policies have impacts across sectors and the combination of policies in each state together shape the overall environment of rights and opportunities of noncitizens even more so than any one set of policies. The accumulated effect creates a further synergy of negative perceptions of immigrants that is difficult to dispel. For example, in states that allow many undocumented residents to pay instate tuition to public colleges, also offer access to a driver's license that allows young adults to better convert their increased access to higher education into better paying jobs that, in turn, will improve their economic well-being, their access to health insurance coverage, and other health-promoting opportunities (Potochnick, 2014;Rhodes et al., 2015). These policies, in turn, help the broader community benefit from an improved tax base resulting from higher levels of employment. Similarly, research suggests that experiences, such as stress due to legal status and living in an anti-immigrant environment, can have a life-long and cumulative negative impact (Torres et al., 2016). The fear of deportation, for example, can have a chilling effect that discourages immigrant women from seeking well-child medical care for their US-born citizen infants (Rhodes et al., 2015;Watson, 2010), which could lead to the delayed detection of illnesses or developmental problems (Yoshikawa et al., 2008). In turn, these problems may have deleterious effects on non-health outcomes, such as lower educational attainment. Thus, different areas of state-level public policy work directly, indirectly, and in synergy with each other to influence the opportunities and experiences of immigrants, producing inequitable conditions across different legal statuses and between states. Immigrant policies may be interrelated with other structural and political factors within states. Research that examines the factors that drive immigrant policy indicate that the creation of policies results from social, economic and political forces within each state (Gulasekaram & Ramakrishnan, 2012;Marquez & Schraufnagel, 2013;Thangasamy, 2010). For example, some evidence suggests that policy makers opt for more restrictive policies in highly politically polarized states where policies that are anti-immigrant may be used as wedge issues by politicians or parties (Gulasekaram & Ramakrishnan, 2012). Other studies suggest that an increase in restrictive policies is associated with demographic changes, although not all states with exclusionary policies are new destination locations with highly visible recent immigration. Studies have variously identified anti-immigrant policies as also being associated with rapid increases in the immigrant or Latino population, with poor economic conditions where immigrants can be accused of competing for jobs with native-born residents, and with large numbers of older voters who may be concerned with social changes associated with immigration (Marquez & Schraufnagel, 2013;Myers, 2007;Thangasamy, 2010). As with the literature on the impact of immigrant policy on the health and welfare of immigrants, this literature on the determinants of immigrant policy tends to focus on predicting single policies or policies in a specific domain. Overall, these policies define the different environments in which immigrants live throughout the nation's states. It is critical to understand the overall policy environment that influences the lives of immigrants, either directly on their health access and decision-making, or more indirectly, as it impacts social and economic structures. 2. Methods for measuring state immigrant policy and assessing its relationship to policy predictors The above framework provides a conceptual approach to assessing the impact of policy at both individual and structural levels and the joint impact of multiple policies. To improve our understanding of the variation in immigrant policy environments across the US, we applied the framework to measure the policy environment in each of the states and the District of Columbia. We then conduct a descriptive analysis to examine patterns in the relationship between these environments and a selection of state characteristics that are both factors that may influence health and that may be associated with trends in immigrant policymaking. --- Policy scan To measure the policy environment across states, we conducted a scan of immigrant policies in the 50 states and the District of Columbia that were in effect by December 31, 2013. We selected a total of 11 policies in the five policy sectors: 1) public health and welfare benefits; 2) higher education; 3) labor and employment policies; 4) drivers' licenses and identification systems; and 5) immigration enforcement. Because of the changes in the structure of immigrant policy-making following PWORA and IIRRIA in 1996, the majority of these policies were enacted after this period. Policies were created through a variety of mechanisms, including legislation, voter initiative, state regulatory action (e.g., university Board of Regents), and court rulings. For the purposes of this review, we focused on the outcome of the policyidentifying whether or not the rights, protections, or eligibility were extended to noncitizens, rather than the specific policy-making mechanism. As described in Table 1, we created indicators that identified an outcome for each policy which allowed us to determine if the policy existed (Yes) or did not exist (No policy) in each state. Each indicator outcome was coded as inclusive (1), exclusive (-1), or neutral (0) depending on its inclusionary or restrictive impact on immigrants' rights to programs and social protections. Policies were determined to be inclusive if they expanded a right or eligibility and exclusive if they restricted a right or eligibility based on immigrant legal status. Neutral impacts were those in which there was an intermediary level between two different restrictive and inclusionary policies and where the lack of a policy did not represent a proactive decision to exclude immigrants (e.g., state opposition to REAL ID is coded as inclusive, while the lack of a legislative resolution on the issue is coded as 0 since in this case no action is not the same as supporting the federal position). We then determined if the policy existed in each state through a systematic review of publicly available policy reports (See Table 1) (NCSL, 2014a;NCSL, 2014b;NCSL, 2014c;NCSL, 2014d;NILC, 2013;NILC, 2014;Schumann, 2004;USDA, 2012). When needed, we verified the existence of a policy using direct searches in online state legislative or regulatory codes. To capture the overall state policy environment, we applied the coding scores to each state and calculated an aggregate score of the policies for each state, with a possible range of +10 to -8. The complete scoring and methodology is described elsewhere (Blinded). The score is an ordinal, rather than interval scale as existing research precludes weighting each policy for varying impacts on health. --- State characteristics data set To understand the extent to which immigrant policy contexts correspond with or are related to other state-level characteristics, we identified key measures of demographic, economic, and political characteristics based on the literature. State demographic and economic characteristics were the percent of the state that was foreign born, Hispanic, non-Hispanic White, over age 65, below 100% federal poverty threshold (FPL), and unemployed, each obtained from the 2014 American Community Survey, and percent undocumented, obtained from the Pew Hispanic Center (Passel & Cohn, 2016). The percent of votes won by the Republican presidential candidate in 2012 (Wooley & Peters, 2016) is used as a proxy for partisanship following the measure used in other research (Gulasekaram & Ramakrishnan, 2012). --- Analysis plan We conducted a descriptive analysis of the relationships between the state policy inclusion score and the state-level characteristics that are related to both social determinants of health and immigrant policymaking. We tested Poisson regression models with demographic, economic, and the political variables to assess the level of their association with the state immigrant policy context. For ease of interpretability we present results as incident rate ratios. --- Results The results show how we applied the immigrant policy framework (Fig. 1) to identify and measure state immigrant policy environments and examine their relationship to other state-level environments. --- State immigrant policy inclusion scores The mean score across all states was -2.5, with the possible range from -8 (a state with all exclusive and no inclusive policies) to 10 (a state with all inclusive and no exclusive policies) (Table 2). Most of the states with high scores (most inclusive) were expected, such as California and Illinois; however, Texas is also among the most inclusive states as a result of several inclusive educational and health policies. New York scored below Texas due to its inclusive health and welfare policies that were countered by mixed policies in other areas such as education. States with the lowest scores (most exclusive) included some that were expected, such as Alabama and Arizona, and some that were not, such as Ohio and Indiana. The state with the lowest score (most exclusive), Ohio, had two labor and employment policies that netted a score slightly above the lowest possible score of -8. --- State immigrant policy and demographic, economic, and political characteristics Table 3 presents the association between the inclusion score and state immigration demographics in (Model A), other demographics (Model B), economic indicators (Model C), and political characteristics (Model D). In the individual models, percent foreign-born (ß=0.07, p <unk> 0.01), percent Hispanic (ß=0.02, p <unk> 0.01), percent unemployed (ß=0.21, p <unk> 0.01), and Republican voting (ß=-0.02, p <unk> 0.01) were each associated with the state level of inclusion. The other state-level characteristics were not associated with inclusion scores in the domainspecific regressions. In the final model (Table 3, Model E), the inclusion score was positively associated with the percent foreign-born and Hispanic in a state and negatively associated with percent Republican-voting. Model incident rate ratios (IRRs) indicate that, an additional percent of the state that was foreign-born (IRR=1.07, p <unk> 0.05) is associated with a 7% increase in the inclusion score. A unit increase in the percent of the state that was Hispanic (IRR=1.02, p <unk> 0.05) was associated with a 2% increase in inclusion score. Finally, a unit increase in the percent of Republican voters in 2012 (IRR=0.98, p <unk> 0.05) was associated with a 2% decrease in inclusion score. --- Discussion Our framework fills gaps in existing conceptualizations of the social determinants of immigrant health. Most frameworks focus on specific macrosocial drivers of immigrant health and social position, such as the intersection of legal status and race (e.g., Asad and Clair), the impact of immigration-related discrimination on health (e.g., Viruell-Fuentes et al., 2007) or structural vulnerability (e.g., Quesada, et al., 2011); these frameworks lay out social processes, but do not identify the actual mechanisms that create and reproduce those vulnerabilities. While a fundamental cause analysis (e.g., Phelan et al., 2010) focuses on how the resources individuals have to maintain health/ treat disease are stratified by a person's position in society (especially by class and race), our framework provides additional insights by highlighting the network of state-level public policies domains that shape the life chances for immigrants in particular. Our framework fits between the broader structural theories (e.g., racism and health) and more local process studies of how institutional practices impact immigrants' health by highlighting the way that governmental action (public policies) directly and indirectly shapes social inclusion and exclusion specifically of immigrant populations by creating a web of policies across institutions and systems that constitute a context of settlement and immigrant incorporation. While a body of research documents how multiple nonhealth policies impact health (Collins & Koplan, 2009), to our knowledge no one to date has brought this type of analysis to focus on the policy domains that shape the life chances specifically of immigrants. While federal immigration law sets the policy stage, state immigrant policies create differential access to rights and opportunities based on an individual's citizenship and legal status. Most research about the health impact of public policies on immigrants have focused on single policies (e.g., Medicaid coverage for undocumented children) or, at best, sectors (e.g., enforcement). Some research limits their analysis to legislatively enacted laws (Wills & Commins, 2018), but we also include administrative actions and court decisions that create policies that are experienced by immigrants, both documented and undocumented. The strength of our approach is that we focus the typical health in all policies framework on the multiple sectors that uniquely shape the context in which immigrants live and their position in each US state based on their citizenship and legal status. State immigrant policies are, therefore, a key component of the social determinants of immigrant Mean score: -2.5; Median score: -3 health. This framework and measures identify the sectors in which states currently have discretion to enact policies and establishes a definition for immigrant policy as public policies that shape the rights, protections, and eligibility of noncitizens. This provides a foundation for future research to assess the variation in specific health outcomes across these distinct policy environments. It can also be used to continue to assess and measure changes in state policy contexts, either as additional states adopt current policies or as new policies emerge. Our policy scan confirms that the nation is made up of varied contexts for noncitizens, as each state has a unique combination of policies. People are exposed to multiple policies simultaneously, making a composite measure of the policy context conceptually more accurate as a contextual factor than any single policy indicator. As the distribution of scores shows, states can be relatively inclusive in some areas (e.g., education), but exclusive in others (e.g., health and welfare). These policies shape the conditions under which immigrants can access services, the level of receptivity to immigrants (or lack thereof) by different institutional resources, the way that these policies and the publicity about them create a social climate that impacts the level of immigrant engagement in their community, and the resultant level of overall stress and vigilance required by immigrants when such policies are not inclusive. In general, the higher the proportion of foreign born and Hispanics in the state, the more inclusive the set of policies. While unemployment is associated with more exclusive policies by itself, this rapidly changing economic indicator was not significant when tested jointly with the political and demographic variables. But these associations are relatively modest. Our findings also indicate that state immigrant policies are subject to unique dynamics, independent of other state factors. For example, Texas is recognized nationally for its restrictive immigrant policies. Simultaneously, it has some inclusive policies for reasons not entirely related to its large Hispanic and immigrant populations (that favor inclusive policies), such as support for unborn child policies that favor prenatal care for all women regardless of legal status (Gray, 2008). In contrast, Ohio does not have a reputation of immigrant hostility, leading us to expect more moderate immigrant policies than its last place showing. While being a swing state in presidential elections, suggesting a moderate electorate, it has a supermajority of Republicans in both legislative houses and a Republican governor (Jacobson, 2013), as well as a relatively small proportions of immigrants and Hispanics. This pattern of findings may seem counterintuitive at first, since exclusive policies are often viewed as a response to changing demographics (Myers, 2007). By 2013, however, the immigration wave of the 1990s had become an established population in many key states (e.g., California and New York) and policies had evolved to be responsive to those large populations that included families with a growing number of voters and businesses that relied on their labor. As a result, the most exclusive policies (e.g., where exclusive policies impact relatively powerless groups that few businesses rely upon) are now often in states with relatively small immigrant populations; Florida and Arizona are notable exceptions where the politics appears to be more important than demographics. These variable trends, and even more so the exceptions to the trends, of inclusion in some states demonstrate the importance of examining multiple policies, and not assuming that a single policy or sector reflects the complete landscape of policies that impact both social determinants, and in turn, health. The independence of inclusive immigrant policies from common economic indicators indicates that the policies are also not simple proxies for other social determinants of health, such as economic climate. The inclusion score is modestly associated with the size of the immigrant and the Hispanic population, and inversely associated with a larger Republican-leaning electorate as expected. Other factors suggested, including a larger elderly population and poor economic conditions, are not associated with the composite score. This does not mean that individual policies are not passed in response to some of these circumstances, but rather that the overall set of policies is independent of a single indicator in these areas. State immigrant policies in the sectors of public health and welfare benefits, higher education, labor and employment, drivers' licenses and identification, and immigration enforcement will continue to change and shape the lives of immigrants. While there have been periods of high and low policy activity around immigrants in the past, the election of President Trump has made federal policy on undocumented immigrants harsher and more exclusive (Kulish et al., 2017), which has in turn led to an increased level of policy activity from states and localities on immigrant benefits and rights. For example, in the last two years, California state policy makers have expanded policies that limit state law enforcement agencies from collaborating with ICE, while Texas passed a bill to mandate this collaboration (McHugh, 2018). In addition, even in inclusive states the increasingly anti-immigrant climate nationally may contribute to immigrants' heightened vigilance and avoidance of available opportunities. For example, the proposal to expand the public programs that count as a "public charge" would make it more difficult for unauthorized immigrants to regularize their status or LPR immigrants to sponsor relatives to immigrate legally. This would be a highly visible and consequential federal policy change that would heighten the national exclusionary context (KFF, 2018). The sweeping impact of that change could motivate new types of state policies and approaches designed to moderate that national policy in ways that protect immigrant health. Our framework can be used to track these types of policy changes and the resulting increase in disparities across states for immigrants in the overall policy contents. The organizations that we cite as the sources of most of our policy scan regularly update their information. Our two primary sources of information regularly update their policy scans: National Immigration Law Center has updated data on drivers' licenses, higher education access, and health care coverage for children and pregnant women (NILC, 2014); the National Conference of State Legislatures publishes a summary of state legislation at the end of every year. Local governments and institutions may also become increasingly important in creating policies for immigrants. Shortly after the election of President Trump, many cities and universities declared or recommitted themselves as "sanctuaries," pledging varying levels of noncompliance with federal efforts to conduct immigration raids or arrests (Njmabadi, 2016;Sanburn, 2016). While sanctuary city policies cannot entirely preclude the presence of enforcement policy, they can mitigate the extent to which ICE is able to collaborate with local law enforcement agencies (Graber & Marquez, 2016). Localities may also play a role in the other policy domains we identified as well, including health care (e.g. through county hospitals), education (e.g. community colleges), identification (e.g. some communities have established local ID's that unauthorized residents can obtain), and employment (e.g. conditions attached to city/county contracts). While we have identified a set of policies that are likely to impact the health of undocumented and other immigrants, more research is needed to determine the extent to which policies, whether inclusive or exclusive, are implemented as written and the extent to which immigrants' lives and actions are impacted by the presence of those policies. For example, reporting undocumented immigrants to Immigration and Customs Enforcement (ICE), if they file wage theft complaints, may be illegal retaliation in some states, but if the action is never prosecuted, then having the policy on the books is only symbolic. Similarly, immigrants may not access services for which they are eligible due to lack of knowledge regarding the services, lack of experience in accessing public services in the U.S., concerns over issues of public charge, and fears of deportation. On the other hand, state policies may be mitigated somewhat by community groups that develop ways to partially work around exclusive policies. For example, most states do not provide state-funded college financial aid to undocumented students who graduated from high school there, but community advocates can help identify alternative funding sources, such as private foundations or individual donors, to support college attendance (MALDEF, 2016). Similarly, there are examples of local advocacy by community organizations and individuals to protect immigrants from enforcement actions, such as a local newspaper reporter in Atlanta who updates information on ICE locations to Facebook and has 250,000 followers (Yee & Vivian, 2017). Studying how immigrants and providers navigate these environments also needs further inquiry to ascertain how mixed policy environments impact social determinants of health, and in turn, health outcomes. We expect that more inclusive policy environments not only promote improved access to medical care, but that both physical and mental health outcomes among immigrants will be better over the longterm. This study has some limitations that should be noted. The inclusion score is designed to show relative inclusiveness, rather than the level of absolute inclusiveness. There are policies within the five areas in our framework that we did not include because of a lack of state-level variation that nevertheless are likely to impact social determinants of health for undocumented immigrants, such as their exclusion from housing subsidies, ineligibility for food stamps, and inability to buy health insurance on the state health exchanges. In addition, because we relied on existing policy and legal sources as our first source of information on policies, there may be policies within the five areas that were not included, such as enforcement policies requiring that

Background: Many conceptual frameworks that touch on immigration and health have been published over the past several years. Most discuss broad social trends or specific immigrant policies, but few address how the policy environment affects the context of settlement and incorporation. Research on the social determinants of health shows how policies across multiple sectors have an impact on health status and health services, but has not yet identified the policies most important for immigrants. Understanding the range and content of state-level policies that impact immigrant populations can focus health in all policies initiatives as well as contextualize future research on immigrant health. Methods: Our framework identifies state-level policies across five different domains that impact the health of immigrants and that vary across states, especially for those without legal status. Our scan shows that immigrants are exposed to different contexts, ranging from relatively inclusive to highly exclusive; a number of states have mixed trends that are more inclusive in some areas, but exclusive in others. Finally, we examine how the relative inclusiveness of state policies are associated with state-level demographic and political characteristics. Results: Contrary to the image that exclusive policies are a reaction to large immigrant populations that may compete for jobs and cultural space, we find that the higher the proportion of foreign born and Hispanics in the state, the more inclusive the set of policies; while the higher the proportion of Republican voters, the less inclusive. Conclusions: Variation across immigrant policies is much larger than the variation in state demographic and political characteristics, suggesting that state-level policies need to be included as a possible independent, contextual effect, when assessing immigrant health outcomes. This policy framework can be particularly useful in bridging our understanding of how large macro processes are connected to the daily lives and health of immigrants.

from enforcement actions, such as a local newspaper reporter in Atlanta who updates information on ICE locations to Facebook and has 250,000 followers (Yee & Vivian, 2017). Studying how immigrants and providers navigate these environments also needs further inquiry to ascertain how mixed policy environments impact social determinants of health, and in turn, health outcomes. We expect that more inclusive policy environments not only promote improved access to medical care, but that both physical and mental health outcomes among immigrants will be better over the longterm. This study has some limitations that should be noted. The inclusion score is designed to show relative inclusiveness, rather than the level of absolute inclusiveness. There are policies within the five areas in our framework that we did not include because of a lack of state-level variation that nevertheless are likely to impact social determinants of health for undocumented immigrants, such as their exclusion from housing subsidies, ineligibility for food stamps, and inability to buy health insurance on the state health exchanges. In addition, because we relied on existing policy and legal sources as our first source of information on policies, there may be policies within the five areas that were not included, such as enforcement policies requiring that police determine legal status when stopping or arresting individuals or labor policies that protect employees from immigration-related employer retaliation. In addition, our scan is limited to policies enacted but without information on the precise scope of the policies or extent of their implementation. For example, the financial threshold for Medicaid eligibility varies by state, yet we only code whether Medicaid was extended to LPRs in their first five years. In some states this will cover a greater proportion of the recent immigrant population than others. Similarly, as noted above, we have no data on how vigorously states enforce their policies. We believe, nevertheless, that our inclusion score appropriately indicates representative policies across the five domains and provides an accurate summary of the variation in the overall policy environment that immigrants will experience. --- Conclusions It is increasingly accepted that health depends on the quality of the environments where people are born, grow up, live, work, and age-the social determinants of health. State-level policies that differentially impact undocumented and other immigrants should be considered a social determinant of health for that population. Our framework highlights policy sectors that have particular relevance for immigrants, their families, and their communities; demonstrates that there is significant variation by state in these policies that are not largely explained by other state-level characteristics; and promotes an intersectoral approach to understanding how immigrant specific contextual factors can mediate broader social determinants of health (e.g., racism, class inequality). This study documents that a significant number of immigrant-health related state policies currently exist across different institutional sectors, that there is significant variation among states in those policies, and the policies cannot be fully represented by the demographic or political characteristics of the state, nor resources available. Researchers need to consider the health impact of a wide variety of nonhealth-related policies and the cumulative impact of those policies on residents who were born abroad, regardless of immigration and legal status, to adequately understand the social determinants of immigrant health. #R01MD012292-02R, and the Department of Health and Human Services, Health Resources and Services Administration, Maternal and Child Health Bureau, Division of Child, Adolescent and Family Health, Adolescent Health Branch Grant #U45MC27709. The contents of this paper are solely the responsibility of the individual authors. --- Declarations of conflict of interest None. --- Declarations of interest None. --- Ethics statement The authors adhere to the ethics in publishing policy. This study involved data on US states and does not include any data on individual persons.

Background: Many conceptual frameworks that touch on immigration and health have been published over the past several years. Most discuss broad social trends or specific immigrant policies, but few address how the policy environment affects the context of settlement and incorporation. Research on the social determinants of health shows how policies across multiple sectors have an impact on health status and health services, but has not yet identified the policies most important for immigrants. Understanding the range and content of state-level policies that impact immigrant populations can focus health in all policies initiatives as well as contextualize future research on immigrant health. Methods: Our framework identifies state-level policies across five different domains that impact the health of immigrants and that vary across states, especially for those without legal status. Our scan shows that immigrants are exposed to different contexts, ranging from relatively inclusive to highly exclusive; a number of states have mixed trends that are more inclusive in some areas, but exclusive in others. Finally, we examine how the relative inclusiveness of state policies are associated with state-level demographic and political characteristics. Results: Contrary to the image that exclusive policies are a reaction to large immigrant populations that may compete for jobs and cultural space, we find that the higher the proportion of foreign born and Hispanics in the state, the more inclusive the set of policies; while the higher the proportion of Republican voters, the less inclusive. Conclusions: Variation across immigrant policies is much larger than the variation in state demographic and political characteristics, suggesting that state-level policies need to be included as a possible independent, contextual effect, when assessing immigrant health outcomes. This policy framework can be particularly useful in bridging our understanding of how large macro processes are connected to the daily lives and health of immigrants.

Introduction In recent years, intellectual standards have increased considerably with the significant development of the Vietnamese economy. According to data from The World Bank, the annual percentage of Gross domestic product (GDP) growth in Vietnam has never been below 4% from 2009 to 2019. Moreover, Vietnam's GDP per capita is rising continuously and strikingly each year. 1 As citizens' incomes and intellectual standards have improved, people are focusing more on their well-being. Furthermore, with around 61.37 millions smartphone users by May 2021 and approximately 68.7% of the country's residents having access to the Internet in 2019, 1,2 numerous impediments to Vietnamese' intention to improve their quality of life have been generally removed. From the researchers' speculation, there might be an increase in the number of Vietnamese citizens who are ready to acquire new technologies to make life safer and more convenient. During the COVID-19 crisis, the country first applied effective measures with strict procedures to tackle this novel virus; 3 however, since the arrival of the delta variant of SARS-CoV-2, several cities in Vietnam have had to enter rigorous social distancing periods. Therefore, it is challenging for residents to go to doctors and receive direct medical care at hospitals as people are afraid of spreading the disease or being infected by the virus. 4 In addition, Vietnam's healthcare system has been suffering from COVID-19 and has become overloaded. Numerous doctors, nurses, and healthcare staff are required to concentrate on the treatment of COVID-19 patients, which also means that other minor diseases might be placed under the priority list. 5 Understanding the context of Vietnam, mHealth might provide a potential solution to COVID-19 and promote healthcare services. According to the Global Observatory for eHealth (GOe) by the World Health Organization (WHO), mHealth or mobile health is defined as "medical and public health practice supported by mobile devices, such as mobile phones, patient monitoring devices, personal digital assistants (PDAs), and other wireless devices". 6 mHealth applications have been widely used in developed countries, enhance the effectiveness of health care services. 7 Specifically, mHealth can be cost-saving, time-saving, and able to provide better reach, increased interaction between patients and doctors by making it easier and faster to exchange information regarding diseases and health conditions. 8 By operating in distanced communication, mHealth applications can also help prevent the contamination of communicable disease within society. 9 Recognizing the benefits and suitability of mHealth in Vietnam, we believe that this application should receive more attention than digital health. First, mHealth services operate on portable devices, making them more convenient for busy residents who often have to go out. Second, it is faster to turn on phones, tablets, or smartwatches than to turn on desktops or televisions; therefore, mHealth is more useful in emergencies. Finally, the prices of mobile phones, tablets, and smartwatches are, in general, considerably lower than those of desktop computers, or other immobile digital devices. Therefore, mHealth applications are becoming more convenient for residents. These reasons motivated us to concentrate on the development of mHealth by encompassing eHealth services. In addition, research on the barriers to and readiness for mobile apps' involvement in healthcare interventions has mostly been conducted in developing countries in Africa, South Asia, and Latin America. 10,11 Moreover, previous studies have used quantitative methods to test and confirm the effectiveness of eHealth interventions. However, before service providers and governments plan to establish mHealth interventions, it is essential to understand whether people accept and adopt these tools. Thus, there should be more qualitative research exploring the factors influencing acceptance of and readiness to adopt eHealth interventions. In addition, research on people's willingness to use mHealth applications has tended to focus on the general population. Meanwhile, the differences in age groups and health status might also play a significant role in affecting the acceptance and readiness of eHealth applications. As per the literature gap above, the purpose of this study is to investigate the factors impacting the readiness, acceptance, and adoption of mHealth applications in Vietnam, a developing country. In this study, the researchers first explored Vietnamese people's level of understanding of mHealth. After assessing the level of necessity and interest of people, the researchers identified the most desired features of promising mHealth applications and elements that could affect the Vietnamese decision to use this type of technology. Moreover, researchers aimed to determine the moderating effects of the COVID-19 pandemic. This study aimed to provide information regarding people's acceptance of mHealth applications to both the Vietnamese government and technological service providers and stakeholders. When they acknowledge consumer demand, they can improve their current application or create a new, effective one. 12 As the theoretical framework of the research, the Unified Theory of Acceptance and Use of Technology (UTAUT) is used to create the interview guide, to understand and analyze the factors that influence Vietnamese acceptance and adoption of mHealth applications in a developing country context. --- The Roles of mHealth in Improving the Healthcare System in the Developing Countries Since the advent of mobile phones, mobile access and the use of mobile phones have become ubiquitous. The significant growth of mobile phones is the penetration of mHealth applications worldwide. 13 People have utilized mHealth technologies as stand-alone intervention tools to help with self-monitoring, patient education, and healthy lifestyle practices, such as better chronic disease self-management. Scholars have conducted various studies to examine the benefits of mHealth applications and have identified gaps in the literature for future improvement. For instance, in light of the rapid virtualization of dentistry, mHealth messages and apps have been considered a viable clinical tool for the enhancement of oral health and the prevention of oral problems. 14 Furthermore, researchers have regarded telemedicine and mHealth as relatively novel intervention modes for controlling systolic blood pressure among stroke survivors, especially those with hypertensive stroke. 15 Telemedicine is also considered to be equally effective for face-to-face care in hepatitis C virus treatment. In general, mHealth solutions have proven to be effective in improving healthcare services. 16 As mHealth applications can provide cost-effective mitigations for prevalent diseases, they can solve financial constraints, which are one of the key obstacles for delivering healthcare services for populations in developing countries. 17 Apart from pre existing problems of childhood infectious, perinatal and nutritional disorders, due to the transformation of life habits and urbanization, the populations of developing countries also face chronic noncommunicable illnesses such as adult cardiovascular disease, diabetes, and depression. 17 However, some of these chronic health conditions can be improved by mHealth technologies. For example, mHealth interventions have been found to improve glycemic control (HbA1c) in patients with diabetes compared to standard care or other non-mHealth approaches. 18 While medical non-adherence is a key barrier for improving the outcome of cardiovascular diseases, mHealth tools can improve medication adherence in patients, 19 and according to previous studies, 20 mHealth applications with a lay support person may yield sustained improvements in antidepressant adherence and depression symptom remission. Therefore, considering the promising effects of mHealth apps, the field of digital health should be researched in developing countries. However, developing countries face many barriers to adopting mHealth services. A systematic review of developing countries identified up to 73 individual barriers with the top three barriers categories namely infrastructure, equipment lacking and technology gap. 11 A systematic review of mobile health adoption for sexual and reproductive health among youth addressed numerous barriers to provider prejudice, stigmatization, discrimination, fear of refusal, lack of privacy, confidentiality, decreased technological literacy, inferior network coverage, and lower linguistic competency. 21 Although the aforementioned research provided valuable insights about mHealth in developing countries, there is no specific research concentrating on the barriers in Vietnam to ensure a more localized and effective app development. Moreover, developing countries are currently facing a new challenge: the COVID-19 pandemic. Hence, we aimed to find a solution for mHealth to promote the population's life conditions and fight the pandemic. --- The Roles of mHealth in Mitigating the COVID-19 Pandemic in the Developing Countries The COVID-19 pandemic has had devastating social, economic and political impacts, making it challenging for governments to act promptly to control the pandemic. 22 Many countries have implemented containment and mitigation measures to prevent the spread of diseases. mHealth technologies have proven to have considerable positive effects in preventing the spread of the COVID-19 pandemic. On a social level, when combined with diagnostic and immune status testing, mHealth technology could be a valuable tool to help mitigate, if not prevent, the next surge of COVID-19 cases. 23 According to a previous study, digital health can contribute to reducing the spread and improving the effectiveness of pandemic control. 22 On an individual level, health apps support users in managing their risk of exposure and symptom progression and help patients to track their symptom development, take medication and maintain a healthy lifestyle. 24 Acknowledging the benefits of mHealth in the current health crisis, multiple studies have investigated the factors affecting its adoption in this pandemic context. Using UTAUT2, one study confirmed that multiple variables such as performance expectancy, effort expectancy, social influence, facilitating condition, and hedonic motivation were significant determinants of mHealth usage intention. 25 Moreover, another study also identified a strong direct positive effect of both situational constraint and health consciousness on the behavioral intention to use mHealth technologies. 26 Second, there have been several papers showing empirical evidence to prove a rising demand for mHealth apps due to the pandemic situation. In the United States, patients with chronic health conditions such as cardiometabolic diseases, respiratory conditions, immune-related conditions, mental health, and obesity are more likely to use COVID-19 mHealth applications. 27 Researchers in China have also identified that parents' willingness to adopt mHealth for their children during the pandemic was relatively high. 28 Considering the benefits of mHealth applications investigated in developed countries during the current situation, the authors are motivated to discover mHealth adoption among developing countries. The researchers were motivated by multiple gaps in the literature regarding mHealth adoption in Vietnam, particularly during the pandemic. First, different countries will require different conditions such as the healthcare environment, status of progress in IT, and economic status which can influence the possibility of introducing telemedicine. 29 However, previous research has only shown that Vietnam has a high possibility of introducing telemedicine services based on facilitating conditions. Therefore, researchers wish to contribute by investigating Vietnamese people's demand, particularly during the COVID-19 pandemic. Second, although patient demand in developing countries during COVID-19 has been widely investigated in Central East Asia and South Asia, there are only two studies on digital health in this pandemic situation in Vietnam but from policy-makers and doctors' perspectives. 22,30 However, the literature has underresearched the desiring features from the users perspective in an mHealth application, specifically to mitigate the current COVID-19 pandemic situation in Vietnam. These gaps motivated the authors to choose Vietnam, a country unfamiliar with digital or mobile health, as the new research context. In Vietnam, mHealth has remained a novel topic both before and during the COVID-19 pandemic. Many product orientation papers published from 2016 to date have examined the practicality, usefulness, and feasibility of some mHealth applications in various fields such as HIV, maternal and infant health, sexual and reproductive health and user satisfaction and acceptance. [31][32][33][34][35] Furthermore, several market-orientation investigations have been conducted. Using a quantitative approach, studies have found that there is a low percentage of mHealth application users; however, the younger generations were mostly satisfied with mHealth innovations that they had experienced and were ready, and felt receptive to more useful mHealth applications in the near future. 36,37 In response to the COVID-19 pandemic, Vietnamese companies have seen skyrocketing needs and interest in the early stages of the situation. 38 However, the majority of mHealth applications are not sufficiently ubiquitous and are unable to reach a large proportion of Vietnamese people. Previous studies have identified that the largest challenges to mHealth sustainability and development in Vietnam are underdeveloped technological infrastructure, lack of attention, and funding from the government. 35,39 However, the context of mHealth in Vietnam has changed considerably nowadays mostly due to COVID-19. The government has devoted large amounts of money to developing mHealth applications such as Bluezone, NCOVI and health information systems and has attracted millions of installations. 40 However, these mHealth COVID-19 applications only hold attention for a significantly short amount of time and have not diversely solved the demands of the residents. In conclusion, we believe that there are significant gaps in the field of mHealth in Vietnam. First, more qualitative studies should be conducted to acquire Vietnamese demand from different dimensions. 22 Second, recent scholars have mostly focused on the younger generation, while the differences in technology receptiveness and understanding or health awareness between age groups and separated health condition groups are also vital problems that more papers need to investigate carefully. 25 Additionally, during the COVID-19 pandemic, there has been a lack of research investigating the demands and viewpoints of vulnerable and limited access to technology citizens such as the elderly or people with chronic diseases, who are sometimes the group of residents requiring the most health care services. For instance, two researchers who concluded that digital health and emerging telehealth technology hold the potential to support Vietnam against the COVID-19 pandemic admitted that their papers came from the perspective of doctors, not residents or patients. 22,30 Third, to build a diversity of mHealth interventions, it is essential to study further mHealth fields, as well as the limited fields that developers believe to be important. Finally, there has been a demand to concentrate on mHealth applications as most of the current interventions studied are based on SMS or phone calling foundations. 11 Acknowledging several gaps, the researchers decided to conduct this investigation to support other authors in solving the unknown. --- The Unified Theory of Acceptance and Use of Technology (UTAUT) In the literature, various theories can explain user adoption behavior, such as Mental Accounting Theory (MAT), Theory of Planned Behavior (TPB), and the Technology Acceptance Model (TAM). 41 Hence there is a need to decide on a theoretical framework suitable for the study goals. Recent literature has employed different versions of The unified theory of acceptance and use of technology (UTAUT) to explore the adoption of mHealth in developing contexts, such as Bangladesh and Iran. 42,43 Under a quantitative approach, these scholars have also found that this theory has high explanatory power regarding the acceptance of mobile apps and wearable technology. 44 Therefore, this study adopts the UTAUT framework to investigate mHealth in the Vietnamese market. According to the literature, 45 UTAUT is a modified version of TPB and TAM that examines user perceptions using four variables: performance expectancy (PE), effort expectancy (EE), social influence (SI) and facilitating conditions (FC). According to this study, while PE depicts how an individual expects that employing technology will improve his or her performance, EE is the degree to which a person expects that using a given technology will not require a significant amount of effort. SI studies how users' families and communities affect the adoption rate of new technological innovations, while FC explores how accessibility to technical support and infrastructure can facilitate the use of technology. The data collected from these variables can guide policymakers and healthcare enterprises in tailoring their products and services to fit market demand and readiness. Finally, while the original UTAUT has four moderators, including gender, age, experience, and voluntariness, the researchers only compared the results between different age groups and the health conditions of the participants and their families to ensure the specificity and concentration of the study scope. Regarding age groups, the existing literature has found that there are different behaviors and motivations to adopt new technology, such as Generation X utilizing mobile phones for information search and Generation Z utilizing mobile phones for entertainment purposes. 46,47 Moreover, it is important to understand the gap between technical knowledge and readiness among Vietnamese citizens to tailor better products and services for each group of participants. 48 Empirical evidence also suggests that the demand for curative medicine and preventive medicine varies between participants with different health conditions. 49 One explanation is that each disease requires treatment. It is common sense that a high perception of health severity can motivate potential users to adopt mHealth technologies. However, one recent study found that health perspectives on perceived susceptibility and perceived severity do not impact the intention to use wearable healthcare technology. 50 To investigate more deeply the conflicts in the research findings, the health condition of participants and their families also plays a key role in the analysis of mHealth products in Vietnam. (See the Study design and data collection for more information.) --- Methods and Methodology Study Design and Data Collection Previous literature has shown the benefits of in-depth interviews and thematic analysis techniques to explore users' intentions and decisions to adopt a new digital product across the healthcare industry in diverse contexts. 51,52 In-depth interviews allow researchers to obtain detailed and valuable information about participants' attitudes, expectations, and concerns. Thematic analysis provides a guideline for organizing collected data and generating practical insights. 53 Hence, this study selected this approach to analyze the motivations and barriers that influence users' intentions and decisions to adopt mHealth products. Snowball sampling was used to recruit 50 participants, including the results of the pilot study (Table 1). The criterion for selection was that the participants were Vietnamese citizens with a date of birth from 1965 to 2005. The researchers categorized the participants into 6 sub-groups based on the age and health status of the participants or their family members. Regarding age groups, the study follows the Pew Research Center's categories: Generation X (1965-1980), Generation Y (1981-1996), and Generation Z (1997-2012). 54 While Generation Z has an age range from 1997 to 2012, the researchers did not acquire data from participants who were born after 2005 to comply with Vietnam's Children Law 2021 on research with minorities. Theoretically, this categorization permits researchers to explore and compare the moderating effects of health condition and age on the UTAUT model. Practically, software developers can use this comparative insight to create personalized products that are more serviceable and fulfil the unique needs of each targeted group. Moreover, the researchers followed the COREQ checklist to ensure comprehensiveness and transparency in the interviews and reporting processes (Additional File 1). 55 Two authors (P.D, H.N) conducted the interviews, conducted the qualitative research training and experience. The two interviewers were one male and one female and were pursuing a Biomed and Medical degree. In the pilot interview, the researchers employed convenience sampling to test and reconstruct the semistructured interview guide (Additional File 2); and in the main study, the researchers did not have formal or informal relationships with most participants. Data was collected from July 2021 to September 2021. All interviews had written records, and some had audio recordings, depending on the participants' permission. The duration of these interviews ranged from 14 to 37 minutes. The researchers determined the sample size based on saturating data; hence, the data collection ended with 50 participants (N=50) because there were no significant and novel data from the additional interviewees. Have chronic conditions or family has chronic conditions 28 (56) The frequency of internet usage Daily 49 (98) 1-6 times a week 0 (0) Less than once a week 1 (2) Never 0 (0) The exposure level to mHealth Experts in fields (mHealth developers, mHealth researchers) 0 (0) Recognize mHealth applications and engage with mHealth daily 6 (12) Engage with mHealth daily, but do not recognize the mHealth applications 18 (36) Never heard of the mHealth concept 26 (52) Before the official meeting, the researchers sent a Qualtrics survey form to ask for participants' written consent and socio-demographic characteristics. In this step, the researchers screened out participants who were unsuitable for the scope of this study, such as non-Vietnamese citizens and those older than 65. If a participant passed the screen round, the interviewers scheduled an interview via mobile phone or Zoom. The in-depth interviews began with an introduction of the research topics and goals to the participants. The interviewers then checked the participants' understanding of the mHealth applications and clarified the important concepts. The interviewers facilitated the questions following the interview guide, which was built based on UTAUT theory. These questions aimed to explore and define Performance Expectation (PE), Effort Expectation (EF), Social Influence (SI), and Facilitating Conditions (FC) in the context of mHealth applications in Vietnam. The interview ended with participants' conversations and thoughts about the research topics. This final step allows researchers to gain additional information beyond the scope of traditional UTAUT theory. --- Data Analysis After data collection, researchers performed data analysis and interpretation. While all data coders have studied qualitative research, two authors, KT and TN, have previous qualitative publications on healthcare and technology adoption. Moreover, all the coders were fluent in both English and Vietnamese. As with the collected data in Vietnamese, the researchers employed a forward and back translation approach when transcribed and translated into English to preserve the clarity and consistency of the information before and after translation. The researchers combined the transcribed data verbatim and notes using Microsoft Word. The research followed the guidelines of the thematic analysis developed in the literature. 57 In which, the researchers coded interviewees' responses manually because the study sample was relatively small and this method ensured nuance and accuracy in finding the distinctions between demands and challenges faced by each group (age and health condition). The researchers familiarized themselves with the transcripts by repeatedly listening to and reading the interview audio or notes. The researchers then constructed a coding tree with precise descriptions of major themes and sub-themes to categorize and classify meaningful patterns and trends into relevant UTAUT constructs. For instance, the initial code "leaner medical/ booking procedure" was placed in the effort expectancy. When new themes or additional information emerges, researchers will work together for sorting, regrouping, or adding new nodes to the coding tree (Table 2). This refinement process has enabled researchers to develop insightful data-driven theories in a concise and punchy manner. The research further sorted these themes and subthemes into a matrix table in Excel to compare the data between age groups and healthcare conditions. After finalizing the data analysis, the researchers used the member checking method to enhance the credibility and validation of the data interpretation and ensure that the participants' thoughts and demands were well understood. 58 Finally, the researchers select 1-2 representative quotations for sub-themes as evidence for the analysis following the procedure suggested by Klappe. 51 Figure 1 shows the relationship between the variables after the data analysis. --- Results --- Social-Demographic of the Interview Participants The data were saturated after the number of participants reached 42, and the interviewers conducted eight more interviews to confirm saturation. The interviewers observed saturation of the data for all three age groups. Moreover, although the researchers firstly categorized "participants with a chronic condition" and "participants whose family member has a chronic condition" into two different categories, they decided to merge these groups due to their similarities in response under UTAUT theory. Hence, 22 participants had no chronic condition, and 28 participants (or their immediate family members) had a chronic condition. The participants' chronic conditions included osteoporosis, gastritis, hepatitis, hypertension, cancer, diabetes, osteoarthritis, and cardiovascular disease. After summarizing the socio-demographic information of the participants in Table 1, the researchers identified some potential biases in the responses, which might not sufficiently represent the general Vietnamese population. Nearly all participants (N=49) used the internet daily. Compared with the average internet user rate in Vietnam of 70% of the population, the study sample may be biased toward higher technological readiness. Moreover, nearly half of the participants did not know the concept of mHealth or its potential applications. Although the interviewers had provided simple explanation for these participants, they only provided comments on technology implementation overall rather than mHealth specifically. Nevertheless, the authors still incorporate these suggestions into the results and discussion because their lack of knowledge illustrates a market gap that policymakers and entrepreneurs must address. On the other hand, four healthcare professionals and eight technology experts provided in-depth strategies on technological adoption and features of the mHealth app. Therefore, this study clearly states whether the quote comes from the perspectives of both experts and customers. --- Awareness and Necessity of mHealth in Vietnam In terms of mHealth awareness, this concept was unheard in nearly half of the interviews. Many participants engaged in mHealth applications daily; however, they did not recognize or be aware of the term. Regarding the necessity of mHealth applications from the interviewees' perspective, most respondents agreed that they were essential in Vietnam. However, each age group and health condition had different levels of necessity and reasons for their decision. Generations X and Y mentioned that their hectic lifestyle prevented them from taking care of their health. Therefore, mHealth interventions are required to support such patients. On the other hand, Generation Z expressed that because current mHealth applications have not fulfilled their demands, they certainly need alternatives. Furthermore, people with family members with chronic diseases have also reported that Vietnamese healthcare systems are overloaded, requiring more help from mHealth. By contrast, a small number of participants said that mHealth --- 272 applications are unnecessary in Vietnam as the country's technology has not been sufficiently developed to make residents rely on them. Therefore, people still prefer face-to-face interactions with medical staff. In addition, some interviewees mentioned that, in general, many Vietnamese citizens do not have sufficient technical skills or knowledge to join mHealth networks. Hence, before deciding to implement any mHealth interventions, the country must first educate its residents, which is a burden on Vietnam's current government. --- Performance Expectancy A one-stop-shop solution was a sub-theme mentioned by most interviewees when asked about their performance expectations for mHealth applications. In this context, a one-stop-shop solution means that the application offers multiple services or products to users. This provides convenience and efficiency to users because they feel demotivated when downloading apps to serve various purposes in daily life. Some highly mentioned features include counting steps, calculating sleep time, and regular checkup reminders to promote healthy lifestyles. In particular, respondents demanded that the providers of mHealth applications partner with various healthcare providers as they desire more straightforward and faster access to the on-site healthcare system. For example, people often have to wait in long queues to receive health checkup results; therefore, they wish to obtain electronic results. In addition, to satisfy different expectations, mHealth applications can be integrated with online booking appointments, contact with nearest hospitals or clinics during emergencies, and a comparison of hospital service quality. With the growing number of people using online services, attendants believe that mHealth applications will gain attention when offering online healthcare consulting, virtual pharmacies and delivery services. During the COVID-19 outbreak in Vietnam, these functions are more pertinent as they reduce waiting time. High-quality information and measurements also play critical roles in the adoption of mHealth applications. All delivered information must be accurate and detailed and must come from reliable sources such as experts, practitioners, and hospitals. I strongly believe one of the main reasons why many people do not use mHealth applications frequently is because the technology nowadays does not provide accurate measurements enough only through mobile phones. (A, technological engineer, Generation X, respondent or family members have chronic health problems) According to most participants or their family members who have chronic diseases, they would strongly prefer mHealth applications capable of measuring the health index and comparing users' index to a healthy state. In addition, many generation X interviewees welcomed articles about herbs beneficial for disease prevention and health enhancement. This is partly because the elderly in Vietnam traditionally favor medicines from wild herbs for mild illnesses. Vietnamese residents tend to self-diagnose their health conditions before seeking medical treatment from doctors. Therefore, respondents also demanded an enhanced medical self-diagnosis experience. They hoped to receive a list of possible medical conditions and corresponding suggestions for self-treatment after entering their symptoms and answering several simple questions. I want when I put information about my health, the application can list out fully what health problem I may have. When the application feels that my symptoms are serious, it can recommend that I go to see a doctor. (B, Generation X, respondent or family members with chronic health problems) One of the most prominent features influencing people's intention to adopt mHealth applications is personalized recommendations with data protection. According to most interviewees, mHealth applications should provide recommendations based on the health status and medical treatment. (I want) the apps to be able to give me instructions to cure my own health problems by myself or caring at home, especially in the current situation (the COVID-19 pandemic). (C, medical student, Generation Z, respondent or family members do not have chronic health problems) Moreover, each time healthcare is sought, apps ask patients for information related to biographical data and health history to make an accurate diagnosis, creating a complicated and time-consuming procedure. If mHealth applications create and store individualized health checkup results then patients and hospitals can save time and effort. Additionally, as suggested by the respondents, personal recommendations also come from studying the daily habits of the users. I want the applications to remind me to do some short exercise after a long period of working time... Or maybe tracking my daily meals to advise better food since I am too busy to even think of what healthy food I should eat. It might seem quite robotic; however, in these modern days, I believe we need such care. --- 274 (D, Generation Y, respondent or family members with no chronic health problems) However, many people emphasize data protection when mHealth applications store personal information. These respondents were afraid that ill-intentioned people possibly try to breach their health-related information such as allergies to serve their purposes. People are also worried about identity fraud, such as credit card numbers, or personal privacy, such as the health status of family members that patients try to hide. Another aspect repeated by many interviewees was the updating of the content. This function mainly revolves around (1) updating the user information and (2) providing information related to the latest health events. Regarding updating user information, interviewees hoped to have their health records automatically updated after each treatment so that they and their doctors could conveniently track their situation. It also includes a quick update of user information changes such as bank accounts, phone numbers and home addresses. For instance, once the users finish their health checks at hospitals, the applications must immediately update and store information related to the health checks for users to easily keep track of. (E, student, generation Z, respondent or family members have chronic health problems) In terms of updating the latest health events, respondents demanded official and accurate news from official sources related to current health issues or breakthroughs in healthcare. Recently, almost everyone has been cared for by the COVID-19 pandemic. Our interviewees also hoped to be informed about vaccination plans, the effectiveness of novel vaccines, or guidance to take care of themselves. --- Effort Expectancy According to most participants, a user-friendly app design is a compulsory feature of mHealth applications. Interviewees from the three age groups and two health condition groups agreed on the critical role of the simple and easy-to-navigate toolbar. Users can find the targeted sections without too many sundry steps, such as hitting too many buttons. Voice search and highlighting warnings or essential information were also two of the most common suggestions for simplifying the interface and improving user experience. Furthermore, many people in the chronic health problems group would like to use applications depicting more illustrations such as symbols, images, or short videos rather than words. While only generation Z suggested that application interfaces should be attractive with neutral colors and eye-catching graphics, only generation X wanted large letters because their vision was aging. Apart from these demands, the youngest generation paid the most attention to irritating advertisements that covered content. Mobile health service creators should concentrate on providing an optimal overall user experience. Most participants preferred applications with simple sign-in or sign-up mechanisms, such as fingerprints and face recognition. If my family found an application which was too complicated to create accounts, we would straightforwardly uninstall the application without even bothering to learn more. (F, music teacher, generation Y, respondent or family members have chronic health problems) Surprisingly, generation X, generation Y, and those with chronic health problems respondents wanted regular notifications on mHealth apps. However, a significant number of generation Z interviewees said that they would love to download apps with fewer notifications because pop-up lines seem annoying and distracting. Several participants from the chronic health problems group and generation Y suggested applying advanced technologies such as Artificial Intelligence (AI) to input user data. With AI, elderly or busy people do not have to upload their information manually. Additionally, developers may want to create multiple methods for users to input data such as taking photos, scanning, talking, drawing, and typing. Moreover, the participants hoped that AI could analyze their data and provide valuable recommendations. Language accessibility was also a concern among the interviewees, especially generation X and Y. Applications were advised to be written in Vietnamese, English, and ethnic minority languages of Vietnam if developers aim to nationwide their services. Most mobile health apps nowadays are developed in English by foreigners so the users have a language barrier and could not take all of the advantage from the applications. (A, technological engineer, Generation X, respondent or family members have chronic health problems) Moreover, it was estimated that nearly half of the respondents desired applications that did not use terminologies while providing concise and understandable definitions of medical conditions. --- Social Influence Endorsements from credible and authoritative sources were the most commonly named social influences on participants' intentions to use an mHealth application. Recommendations from doctors, healthcare experts were the most reliable sources of information mentioned for each group of respondents. Moreover, encouragement from the authority to use an application was frequently shared as many respondents named Bluezone -an application widely introduced by the Vietnamese government. Respondents with family members with chronic health conditions tended to be influenced by applications created by large medical organizations or companies. However, respondents whose family members did not have chronic health conditions saw newspapers as credible sources for the introduced application. Word-of-mouth recommendations were the second-most frequently mentioned sub-theme across every group of respondents. The introduction of friends was also widely claimed by respondents to be an effective source of introduction. They believe that their friends' recommendations will be effective and reliable since they have already adopted them. Similarly, suggestions from family members and relatives were strongly trusted as respondents believed: If my relatives recommend an application, definitely it is useful because they truly care for me... I will install it when I have time. (G, Generation X, respondent or family members with chronic health problems) Perceived quality on a digital platform is the impression of excellence that a customer experiences regarding a product through advertisements on a digital platform. Compared to generation X and generation Y, generation Z tended to use a product when it was widely and constantly advertised on social media, especially Facebook, which is the largest social media in Vietnam. 59 In addition, respondents from both health groups were more willing to install an application if it was introduced on Facebook pages and groups about health. As one of the most common forms of entertainment in Vietnam, television advertisements and programs are commonly mentioned as an effective source of introduction, influencing respondents' decisions to use mHealth applications. --- Facilitating Conditions Careful and friendly customer service plays an important role in people's intentions to use mHealth applications. The respondents hoped to have clear and flexible instructions to familiarize themselves with the application. Moreover, an mHealth application should provide direct guidance via hotlines or chats to save time and ease of effort, especially for users who are unaccustomed to mobile phones. I want to have someone to tell me to touch which button because I am afraid that if I touch randomly, my application will crash. (H, Generation X, who or their family members have chronic health problems) In addition, the respondents valued an application provider who had quick replies to questions or feedback from users and showed a willingness to improve. Furthermore, some generation Y and Z participants required the application to have a support fan page on Facebook or a website providing instant help or important news. Technical requirements were also a major concern for the respondents when installing and using an application. Generation Z respondents demanded that mHealth applications take up small storage spaces on their devices because they needed to install numerous other applications. This group of respondents also required applications to operate smoothly without crashes. The respondents also mentioned the application's compatibility with several different operating systems and technological products, especially generation X participants. Additionally, the respondents also paid attention to whether the applications were battery consuming

Background: Vietnam's economy and intellectual standards have witnessed significant development, improving conditions for residents to acquire novel mHealth applications. Additionally, the outbreak of the COVID-19 pandemic has influenced Vietnamese awareness of healthcare; however, previous studies have only been clinician-centered rather than customer-centered. Methods: This study addresses this literature gap by interviewing 50 Vietnamese participants grouped by age, namely Generation X, Generation Y, and Generation Z, and health conditions, namely whether participants or family members have chronic illness. The study utilized semi-structured and in-depth interviews to collect the data and used thematic analysis to analyze the data under the unified theory of acceptance and use of technology framework. Results: Most participants were willing to adopt this technology and demanded a convenient and user-friendly one-stop-shop solution, endorsements from credible and authoritative sources, and professional customer services. However, each group also had distinctive demands and behaviors.This study contributes theoretically by providing context-rich demand for Vietnamese customers across three generations and healthcare conditions during the COVID-19 pandemic and comparing their behavior with pre-COVID literature. While this research provides helpful information for potential app developers, this study also suggests that mHealth developers and policymakers should pay more attention to the differences in the demand of age groups and health conditions.

respondents hoped to have clear and flexible instructions to familiarize themselves with the application. Moreover, an mHealth application should provide direct guidance via hotlines or chats to save time and ease of effort, especially for users who are unaccustomed to mobile phones. I want to have someone to tell me to touch which button because I am afraid that if I touch randomly, my application will crash. (H, Generation X, who or their family members have chronic health problems) In addition, the respondents valued an application provider who had quick replies to questions or feedback from users and showed a willingness to improve. Furthermore, some generation Y and Z participants required the application to have a support fan page on Facebook or a website providing instant help or important news. Technical requirements were also a major concern for the respondents when installing and using an application. Generation Z respondents demanded that mHealth applications take up small storage spaces on their devices because they needed to install numerous other applications. This group of respondents also required applications to operate smoothly without crashes. The respondents also mentioned the application's compatibility with several different operating systems and technological products, especially generation X participants. Additionally, the respondents also paid attention to whether the applications were battery consuming as they frequently used their devices for other purposes. --- 276 Fee and payment method was a factor influencing respondents' intention to use an application. This is similar to Price Value as a separate variable in the Unified Theory of Acceptance and Use of Technology Extension Model 2. 60 However, the authors categorized this sub-theme as a factor in the Facilitating Condition. The application price was an essential element affecting many respondents as they wanted to have a free application. Generation Y respondents frequently mentioned fees as a significant concern and required low or reasonable prices for an mHealth application compared with other generations. This concern poses a significant barrier for investors and stakeholders as respondents do not usually pay for an application but prefer an application with no advertisements. Multiple respondents, especially generation Z, required mHealth applications to provide diverse payment methods because many do not possess internet banking services. Since currently I do not have my own bank accounts, paying and transacting (for the mHealth services) is pretty problematic for me... I think pay by cash (will be easier). (I, high school student, Generation Z, respondent or family members have chronic health problems) --- The Impacts of COVID-19 on the Behavior of Vietnamese Users Toward mHealth As mentioned in the literature review, this research was conducted during the COVID-19 pandemic; therefore, residents' demands and attention toward mHealth applications have significant features related to this disease. First, nearly all the respondents mentioned COVID-19 once or twice during the interviews. Some participants shared that they believed mHealth applications would be beneficial in this situation and wished to take advantage of them as soon as possible. Respondents mentioned a desire for app functions related to COVID-19, such as getting official COVID-19 news, managing health declaration and vaccination plans, and providing guidance for self-care. Moreover, because respondents can diversely provide their demands for COVID-19 related apps, they pay significant attention to the problem. In general, the COVID-19 pandemic has made people care more about their health status and induce mHealth applications among people. --- The Roles of Age on the Behavior of Vietnamese Users Toward mHealth Overall, age significantly contributes to people's intention to use an mHealth application. First, people of different age groups play different roles. Generation Y included people who raised a family, had young children, had old parents, and were primary breadwinners. Therefore, many researchers have focused on application fees and family health tracking services. Furthermore, since generation Y is the primary reproductive age in Vietnamese society, only this age group demanded sex-related applications that protect them from sexually transmitted diseases or birth control. In addition, generation Z is the age of many students and young adults. The researchers found that most people who suggested entertaining content within applications or being influenced by social media were generation Z. Second, we recognized that appearance awareness was also the main difference between generation Z and generation X. The younger generation appeared to focus more on applications that help them improve their skin and weight or make them look more attractive, healthy, and bright. However, the older generation mainly focuses on forming healthy lifestyle applications that improve their health conditions or diseases. Third, physical health condition differences between age groups were believed to influence the Vietnamese decision to use mHealth applications. Generation X is the group of citizens who might have many health problems due to their older age; therefore, they face the highest risk of being severely affected by COVID-19. Coincidently, this was also the group that concentrated on COVID-19-related applications. Although generation Y also cared about forming healthy lifestyles, they concentrated on preventing possible diseases since their health conditions were still healthy. Finally, while technological proficiency among the three age groups was significantly different, the researchers also observed that only generation Y and Z desired applications with advanced technologies such as AI. Only generation X demanded multiple guidance methods and several needs for instructions. Therefore, technological proficiency is a factor that should be considered. --- The Roles of Health Conditions on the Behavior of Vietnamese Users Toward mHealth This research discovered that health conditions make a big difference between residents' choices and judgments of mHealth applications. People who themselves or their family members possess chronic health problems appeared to care more about mHealth applications. Furthermore, only this group of interviewees provided suggestions about providing services for particular diseases such as diabetes, blood-pressure diseases, and osteoarthritis. A generation Z student whose father had high blood pressure disease and a generation Y busy piano teacher whose mother had diabetes mentioned the demand for mHealth. They would be very appreciative and feel secure if they could use mHealth applications to control chronic health problems of their parents and notice them if anything happens. Specifically, the chronic health problems group was the only health condition group that proposed more accurate measurement services and online pharmacies with detailed information. This might be because they have to engage with medical activities frequently. Finally, people with chronic diseases also require SOS services with single-touch mechanisms for mHealth applications and express the demands of applying AI. In contrast, people with normal health conditions only provide or require general and basic demands or technologies. It is also important to note that some interviewees in the healthy group admitted that they did not have enough health awareness since they were still fine after all. --- Suggestions from Participate to Promote and Retain Customers for mHealth Applications In our research, the researchers also asked respondents' suggestions to promote mHealth applications. Apart from the aforementioned factors in Social Influence that can prompt the user's intention to install a new mHealth application, there are many forms of introduction that mHealth service providers can attract users. Several respondents included advertisements from Key Opinion Leader (KOL) and artistic celebrities. Usually with different groups of people, they will follow different (Youtube) channels. For me, I usually watch videos by Youtubers with medical content. If it (the promotion) comes from people who have experienced (using) the application, people will try to use it. (C, medical student, Generation Z, respondent or family members do not have chronic health problems) In addition, some respondents from generations X and Y recommended that an mHealth application be advertised by celebrities relating to health or medicine; otherwise, people will be skeptical about the quality. In addition, mHealth service providers should have campaigned for young people to go house-to-house to reach and introduce to the elderly. Moreover, mHealth services can also advertise on billboards on the roads as Vietnamese people usually ride motorbikes slowly on the streets. Further, mHealth providers can also place devices on roads for people to give mHeath services a try. In case the users need to pay for the application fee, developers should offer a free trial with access to most of the features to fully test the application. Additionally, Generation Y respondents also suggested introducing mHealth applications in large health events. Since most of them are students for generation Z, they mentioned school as a reliable source of recommendation to them. Due to the culture of non-paid applications in Vietnam, the respondents did not prefer to pay for mHealth applications directly. However, most respondents were willing to have application fees covered by social health insurance, even though the insurance price may rise. In contrast, a few generation Y respondents rejected this idea and wanted to pay for a monthly subscription. I want to pay subscription monthly because for example in current pandemic situation, I am not going to work so I do not pay (for the social health insurance), so I can not use the applications although this is a critical health crisis. (J, Generation Y, respondent or family members do not have chronic health problem) The researchers also received several suggestions to increase users' engagement with mHealth applications. Most of the respondents emphasised personal experience when adhering to the application rather than only social influence. Many --- 278 interviewees also placed the suitability of the mHealth application to fit users' demands as an essential factor determining the engagement of the users. If the application fits with my demands I will definitely use it and I believe other people may do so. (H, Generation Y, respondent or family members have chronic health problems) Due to the fact that people might get bored or lazy quickly after a particular amount of time using the applications, some respondents proposed that applications should give users scores for frequent access. Moreover, taking Duolingo as an example, mHealth applications should also have cute, sulked and trendy reminders or notifications to attract the users to open the application. --- Discussion The results are summarized and visualized on Figure 1. Regarding people's awareness about mHealth, most participants are not aware of this application, although many of them use some mHealth applications daily. However, the majority of respondents agreed that they need the support of mHealth applications for a more convenient life. In terms of Performance Expectancy, interviewees demanded a one-stop-shop solution that offers multiple services or products to users at the same time. They hope for a precise measurement of the health index and detailed information for a better self-diagnosis experience. Moreover, the application should make personalized recommendations and possible health risk warnings but still ensure privacy protection. With Effort Expectancy, potential users prefer user-friendly app design such as more illustrations, fewer words, and use easy-to-understand language, especially for generation Y and generation X. With Social Influence, the three important factors are word-of-mouth recommendations, endorsement from credible and authoritative sources, and perceived quality from digital platforms. With Facilitating Condition, this study identifies the importance of direct guidance via hotline or chats and social media on mHealth adoption. On the technical side, the customers prefer that the apps do not consume battery or large space on the phone memory. On the financial side, they want diverse payment options and the fee is covered by insurance. In 2018, a previous study only researched the availability and potential future of the mHealth industry in Vietnam based on customer demands. 35 In 2021, studies researched the role of mHealth in mitigating the COVID-19 impacts only from medical practitioners and policy perspectives. 22,30 While these studies have concluded the benefits of mHealth, their authors acknowledged the need to research Vietnamese users' acceptance and readiness as a future research direction. Since the study had a small sample size, the study findings cannot be generalized to the Vietnamese population. However, this study's findings shed light on the empirical gap proposed by the international literature. This study re-confirms friendly visualization and easy-to-navigate interfaces as important dimensions of effort expectancy. Regarding social influence, credible and authoritative sources and word of mouth from close networks also contribute to the adoption of mHealth services. These findings are indifferent to pre-COVID-19 empirical studies in Vietnam and other developing countries. 36,61 Hence, practitioners may invest in the user interface and experience to improve service delivery and adoption of mHealth in Vietnam and concentrate on social branding to attract potential customers. The advertisement should concentrate on the ease of user-app interaction and seek partnerships with well-known medical influencers in social media or recognition and certification by government departments. 62 As these demands and behaviors are not affected by the COVID-19 situation, the authors suspect that this strategy might be transferable to the post-COVID-19 situation. Since the current study only gathers customers' opinions, the authors encourage future research to conduct a cost-benefit evaluation or structural equation modeling (SEM) to confirm and compare the effectiveness of our suggestions on customers' behavioral intentions and customer loyalty. This study acknowledges new demands for mHealth services compared to previous studies. First, the author acknowledges the importance of language accessibility to ethnic groups. While the scoping study only recognized language barriers as a minor theme in their analysis, the author argues that this sub-theme must not be overlooked. 35 According to the Development Policy Center of The Australian National University and United Nations Vietnam, minority and migrant groups are facing a disproportionate impact caused by the COVID-19 pandemic. 63,64 As these citizens usually lack access to healthcare facilities and even the internet, developers should remove the language barriers to adopting mHealth. The authors encourage mHealth developers, the government, and policymakers to provide minority languages in their app interfaces. This consideration also follows the Vietnamese government motto: "To leave no one behind." Second, the scoping study mentioned that none of the participants in the 2018 study mentioned the need for privacy and confidentiality in health services. 35 However, in 2021, two studies mentioned that Vietnamese users value patient privacy, but this demand is difficult to achieve with the current technology. 22,30 Our study confirms the demand for privacy in the context of mHealth. Moreover, our study participants also emphasized the convenience of app usage, including the login and registration processes. The app can demotivate users if login and registration require a lengthy process, such as two-layer passwords or compulsory reading of data rights and data collections. Because users are not keen on layers of security or read data collection and usage policies, the demand for convenience and privacy creates a paradox in app development. Two potential solutions were proposed. During the open conversation, some participants suggested fingerprint or iris authentication, which has become a primary authentication process in smartphones. This process can shorten login time and ensure user authentication. Second, app developers should create short and easy-tocomprehend consent forms to collect unclassified health data from the users. Future research should investigate the design of shortened and simplified consent forms and the technical effectiveness of these methods. During this COVID-19 pandemic, most participants look for an app that can provide personalized treatments or healthcare advice as access to physical healthcare is limited. The authors also recognized another paradox regarding user demands. While users are reluctant to allow the mHealth application to process their healthcare data and are not active in updating their medical information, they demand personalized and up-to-date information from the internet. Because the authors of this study are not Artificial Intelligence or Computer Science experts, this study encourages future researchers to find technical solutions or business models that can comply with the dual standards of users. Meanwhile, quantitative research can be conducted to investigate the extent to which Vietnamese users value personalized advice concerning their privacy. This data-driven research may guide app developers, policymakers, and entrepreneurs to tailor mHealth apps and develop effective mHealth launch strategies. By separating the collected data into age groups and health conditions, this study adds new information to the current literature. First, the study recognizes that Gen Y are the most potential user because they are responsible for caring for their children and elders in the family. They are also financially and technologically ready to adopt the services. This finding supports the previous research in developing country context. 65 Second, the study also sees that Gen Y and Gen X pay great attention to the perceived severity of their health; hence, they are more likely to adopt mHealth service. These groups also demand more convenience and ease-to-use elements of the app. However, the younger group demands hedonic values, such as gamify mHealth services, to attract their adoption. This finding also supports the previous metaanalysis in China. 66 Since the previous studies are researching a quantitative approach, they cannot fully explore and customize the mHealth service according to the demands of users. Our study provides a context-rich comparison and analysis demands on each group. Moreover, since the behavior of Vietnamese users is similar to that of China and Bangladesh, Vietnamese practitioners may seek advice and follow successful mHealth practices from these countries. Still, future research in Vietnam should employ quantitative tests to confirm the behavior of each age group. This study also responds to the recommendation for further research about the apps that connect doctors and patients from patients' perspectives. 30 Our study shows that while mHealth services for treatment are demanded more by Gen Y, Generation X, and people with chronic health conditions, all participants are interested in adopting mHealth solutions. Unlike during pre-COVID-19 time, where study recognized the mHealth apps were designed to target vulnerable and hard-to-reach populations, 35 future mHealth apps must account for the behaviors and demands of all populations as the COVID-19 pandemic impacts upon them. Moreover, during 2018, they also recognize that each mHealth app typically serves a different purpose, such as raising awareness and monitoring health. 30 This development is suitable for the pre-COVID-19 situations because the demand for healthcare services and healthcare information is lesser than the current time. In contrast, most participants in our studies demand a more "one-stop-shop solution" with a strong interest in health indicator measurements, personalized medical advice, and up-to-date content. Future researchers and app developers should pay attention to this new demand change in Vietnam. --- Conclusions This study employs qualitative methods to investigate the acceptability and readiness of Vietnamese residents towards mHealth application under the UTAUT framework. All participants show interest in adopting mHealth applications and demand a comprehensive health solution to protect them from the COVID-19. The study also compared the behaviors between different age groups. The most important remark is that while most Gen X and Gen Y participants demand mHealth applications to prevent the severity of their health, Gen Z demands more vivid, sociable, and gamify products. Concerning the theoretical contribution, this research highlights a context-rich demand from Vietnamese customers across three generations and different healthcare conditions. It provides a first step for future researchers to formulate a conceptual model that targets the needs and demands of each different group. This study also addresses the gap identified by the recent literature on mHealth. First, it compares and contrasts how users' behaviors in adopting mHealth applications differ before and during the COVID-19 pandemic. Second, the study addresses the acceptability of mHealth in Vietnam, answering the limitations of previous studies. This finding can be combined with previous studies to aid antipandemic policies in Vietnam, and develop mHealth solutions. This study also recognizes and proposes two paradoxical demands of Vietnamese users and the barriers to adopting mHealth services, providing directions for future marketing strategies and tech innovations. On the practical contribution, this study suggests that mHealth developers and policymakers pay more attention to the minority ethnic group and the differences in demand of age groups and health conditions. The study observed the strong roles of family recommendations and government endorsed on the users' adoption of mHealth service. Hence, the mHealth developers should collaborate more closely with the Vietnamese government, medical experts, and key opinion leaders on healthcare to build trust and encourage the adoption of mHealth products. Moreover, Gen Y participants have responsibility for their family health; therefore, the mHealth providers and government can target them first to increase the adoption of mHealth solutions in Vietnam. This study also has limitations that future research should address. Firstly, although this research has reached saturation at all sub-group levels, the sample size of 50 is insufficient to represent the demands, acceptability, and acceptance of all Vietnamese populations. For instance, due to the COVID-19 pandemic, all of the interviews are conducted online or via telephone and have financial stability; hence, the study cannot study the demands and needs of Vietnamese residents who do not have access to the Internet and mobile devices. Subsequent studies should concentrate on a larger and more diverse population. Since the research is conducted via qualitative methods, the authors cannot compare the cost-benefit effectiveness of each suggestion or rank between different demands of participants. Finally, this study views Vietnamese residents' demand, acceptability, and readiness from the UTAUT framework; hence, it oversees many variables that might influence the users' adoption intention, such as trust or financial contribution. Future research can account for other models to develop a more comprehensive view of the topic. --- Data Sharing Statement The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the privacy of participants. The language of the data is Vietnamese. --- Supplementary Material The authors have provided the question guide (Figure S1) to increase the transparency in data collection and data analysis. --- Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki. Ethical review and approval were waived for this study. Under Circular No.4/TT-BYT 2020, the scope of IRB is only covering Biomed studies (defined in the Circular at Article 15, Section a: (1) clinical trials of drugs, (2) equipment, and (3) other products that have not yet been licensed for circulation in Vietnam). Because this study was conducted for information gathering for service development from the general population, the ethics committee review was not required. --- Informed Consent Statement Written informed consent has been obtained from the participants to publish this paper from their survey filling. The researchers also have collected in written and recorded verbal agreement that the participants have agreed to the publication of anonymized responses. --- Consent for Publication This manuscript does not contain any identifiable data. --- Author Contributions All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work. The authors made equal contributions. The sequence of authors in the author list is in the alphabetical order of their first names. --- Disclosure The authors declare no conflicts of interest for this work.

Background: Vietnam's economy and intellectual standards have witnessed significant development, improving conditions for residents to acquire novel mHealth applications. Additionally, the outbreak of the COVID-19 pandemic has influenced Vietnamese awareness of healthcare; however, previous studies have only been clinician-centered rather than customer-centered. Methods: This study addresses this literature gap by interviewing 50 Vietnamese participants grouped by age, namely Generation X, Generation Y, and Generation Z, and health conditions, namely whether participants or family members have chronic illness. The study utilized semi-structured and in-depth interviews to collect the data and used thematic analysis to analyze the data under the unified theory of acceptance and use of technology framework. Results: Most participants were willing to adopt this technology and demanded a convenient and user-friendly one-stop-shop solution, endorsements from credible and authoritative sources, and professional customer services. However, each group also had distinctive demands and behaviors.This study contributes theoretically by providing context-rich demand for Vietnamese customers across three generations and healthcare conditions during the COVID-19 pandemic and comparing their behavior with pre-COVID literature. While this research provides helpful information for potential app developers, this study also suggests that mHealth developers and policymakers should pay more attention to the differences in the demand of age groups and health conditions.

Background An estimated 358,000 maternal deaths still occur worldwide each year [1]. Most of these deaths occur during labor, delivery, or the first 24 hours postpartum, and most complications cannot be predicted or prevented [2]. Where women deliver, who attends them, and how quickly they can be transported to referral-level care are thus crucial factors in determining the ability to successfully intervene [3]. The general pattern of utilization of maternal health services contributes to the incidence of maternal mortality and morbidity [4]. However, many women do not have access to the services they need for various reasons. Moreover, socio-economic factors, cultural traits, and local customs are associated with health-seeking behavior. For instance, there are places where services are available but women do not make use of them [3]. Women may prefer to stay at home for a delivery so that they can take care of family members such as young children or elderly relatives and manage their daily household chores [5]. Previous studies have shown that several factors are predictors of women accessing health services [6,7]. It is important to identify which factors lead women to deliver at health facilities. Social support is also an important variable influencing health [8]. However, little attention has been paid to the possible association between social support and place of delivery. The aim of the present study was to determine the association between social support and place of delivery, and other factors influencing delivery location. --- Place of delivery and types of birth attendant in Kenya Delivery with skilled birth attendants is critical to the reduction of maternal mortality [3]. In Kenya, 43.8% of laboring women were assisted by skilled birth attendants, and 42.6% delivered at health facilities between 2008 and 2009 [9]. The place of delivery, for example, health facility or home, is of great significance: home births are associated with increased maternal mortality, since women delivering at home do not have access to the professional assistance provided in facilities. In Kenya, the National Reproductive Health Policy (2007) has brought about a paradigm shift towards a focus on skilled birth attendants for all pregnant women, thus necessitating a policy change regarding traditional birth attendants (TBAs) as providers of delivery services [10]. However, TBAs continue to play a vital role in delivery, assisting with 27.6% of births. Moreover, relatives and friends assist with 21.2% of deliveries between 2008 and 2009 [9]. In Kericho, the location of the present study, the proportion of women delivering at a health facility remained at 50% in 2007 [11], a proportion considered inadequate. --- Social support Social support from a spouse or partner and a social network of family and friends has the potential to influence women's decisions regarding obtaining prenatal care [12]. As in most societies, women have traditionally relied on other women for social support during pregnancy, childbirth, and breastfeeding [13,14]. Female relatives and friends accompany laboring women to maternity units [15], and the presence of a female relative during labor is associated with improved labor outcomes [16]. Social support has been measured in numerous ways. One frequent criticism of research in this area is the lack of consensus about social support in terms of its definition and how best to measure it [17][18][19]. House, a sociologist, specified that potential forms of social support were emotional support, appraisal support, informational support, and instrumental support [18]. Schaefer classified social support into tangible, emotional, and informational support, and he focused on relationships between these types of social support [19]. A previous study found that social support is related to health behaviors [12], and absence of social support was associated with increased maternal mortality [20]. To obtain a better understanding of the effect of social support on place of delivery, the source and types of social support should be considered. --- The Kenyan context Maternal mortality in Kenya remains high at 488 per 100,000 live births in 2008-09 [9]. Kericho District (comprising East and West Kericho) is located about 260 km from the capital, Nairobi. Kericho District has a population of 503,468 [21]. This district had nine hospitals, 11 health centers, five health clinics, one maternity nursing home, and 97 dispensaries [11]. The doctor/patient ratio was 1:15,000, and 40% of houses were located within 15 km from health facilities [21]. --- Methods --- Subjects This cross-sectional study was conducted from September to November 2011 at Sosiot Health Center, Kericho West, Kenya. Information was obtained by means of a structured interview using a questionnaire. Kericho was one of the target districts of two recent community health projects aimed at improving maternal and neonatal health. Those projects were implemented by the Kenyan Government in collaboration with the Japan International Cooperation Agency (JICA) and a Japanese non-profit organization, Health and Development Service (HANDS). Promotion of delivery at health facilities was included in the message of the one of those projects [11]. Sosiot Health Center was involved in both of these projects. The center is situated beside a major thoroughfare, and is a 15 to 20 minute car journey from the center of commerce in Kericho. The target population of this health center was 43,493 people [11], making it a larger scale center than other health facilities in this region. --- Questionnaire The respondents were 306 mothers aged 18 to 49, who brought their babies to Sosiot Health Center for immunization within their first year of life, from September to November 2011. We excluded from the analysis two respondents who delivered on the way to the health facility and one respondent who did not answer the question on marital status. Thus data from 303 respondents (99%) were analyzed. Local people who were at least high school graduates and spoke English, Kiswahili (the national language), and Kipsigis (the dominant language spoken at the study site) fluently were hired as research assistants. They were trained by the researcher so that they conducted interviews smoothly and protected respondents' privacy. The interviews were conducted at Sosiot Health Center. The questionnaire was constructed to ascertain the demographic characteristics of the subjects, as well as their experiences of delivery including place of delivery and social support (see Additional files 1, 2 and 3). --- Dependent variable The primary dependent variable was the place of delivery of the latest child, dichotomized as health facility or home. Health facility deliveries included those at the dispensary, health center, sub-district hospital, district hospital, national hospital, and private hospital/clinic. --- Demographic and birth experience variables We collected data relating to maternal age, infant age, tribe, education level, occupation, marital status, economic status, medical insurance, the time required to reach the nearest delivery facility, types of birth attendant for home delivery, and parity. Information on household assets (clock or watch, electricity, radio, television, mobile telephone, non-mobile telephone, refrigerator, and solar panel) was used to derive a wealth index using the Kenya DHS [9]. Respondents were categorized into three levels by the wealth index. The time required to travel to the nearest delivery facility was calculated in terms of total minutes by foot, bike, shared taxi, and private car or taxi. Respondents were then divided into two groups by the median travel time. --- Social support variables Social support was measured in terms of: 1) Support for daily tasks (housework, fetching water, and farming), and 2) Advice to deliver in a health facility. Support for daily tasks indicated instrumental support, and advice regarding facility delivery reflected informational support. 1) Respondents were asked if they had support for housework, fetching water, and farming, and by whom. They chose from the following answers: husband, mother-in-law, mother, father-in-law, father, sisters-inlaw, sisters, brothers-in-law, brothers, female relatives, children, friends, neighbors, domestic servants, co-wives, and others. 2) Respondents were asked whether somebody had advised them to deliver at a health facility. They chose from the following answers: husband, mother-in-law, mother, father-in-law, father, sisters-in-law, sisters, female relatives, children, friends, neighbors, domestic servants, co-wives, health staff, and others. --- Ethical considerations All women who participated provided written informed consent after reading through the consent form with the interviewer. They were informed that they had the right to refuse participation or to withdraw from this study at any time without prejudice to themselves. This study was approved by the ethics committees of Nagasaki University (Nagasaki, Japan) in December 2010 and Kenya Medical Research Institute (Nairobi, Kenya) in July 2011. --- Data analysis The chi-square test was used for nominal scale data, whereas the Cochran-Armitage trend test was used for ordinal scale data. The simultaneous effects of factors on facility delivery were analyzed using linear logistic models. The following factors were included in stepwise logistic regression: age, education level, occupation, economic status, medical insurance, the time required to travel to the nearest delivery facility, support of motherin-law for household work, support of husband for farming, support of sisters-in-law for fetching water, advice on facility delivery from family or neighbors, advice on facility delivery from health staff, and parity. Odds ratios (OR) with 95% confidence intervals (CI) were calculated. IBM SPSS Statistics Version 21 was used for statistical analysis (SPSS Inc, Chicago, IL, USA). --- Results Socio-demographic characteristics and birth experience are shown in Table 1. Among the 303 respondents, age ranged from 18 to 41 years and median age was 23 years. Around 86% of respondents had some primary education. The most common occupation was farming. About 97% of women were Kipsigis, and more than 80% of respondents were married. In their most recent delivery, 58 (19.1%) women delivered at home and 245 (80.9%) delivered at health facilities. Just less than half of women who delivered at home were assisted by their mother-in-law, and the next largest proportion was assisted by neighbors. Table 2 presents the associations between sociodemographic characteristics and place of delivery. Bivariate analysis indicated that facility delivery was more likely in younger women (P = 0.002), unmarried women (P = 0.008), primiparae (P <unk> 0.001) and those with financial capability or high educational attainment (P <unk> 0.001). Table 3 shows associations between support for daily tasks and place of delivery among married women. Married women whose husbands supported them in farming and whose neighbors helped them fetching water were less likely to deliver at health facilities (P = 0.003 and P = 0.021, respectively) than those whose husbands and neighbors did not provide this support. Table 4 shows associations between support for daily tasks and place of delivery among unmarried women. Unmarried women whose mothers supported them in housework and those whose sisters helped them fetch water were significantly more likely to deliver at health facilities (P = 0.002, P = 0.042, respectively) than those whose mothers and sisters did not provide this support. Table 5 shows the associations between advice regarding facility delivery and place of delivery among married women. Married women who were advised by their mothers-in-law or by health staff to deliver at a health facility were more likely to do so (P = 0.015, P = 0.022, respectively) than those who were not. Table 6 shows the associations between advice regarding facility delivery and place of delivery among unmarried women. There were no statistically significant differences between those who were advised to deliver at a health facility and those who were not. Table 7 summarizes the results of multivariate logistic regression analysis of factors associated with the place of delivery in married women. Facility delivery was more likely among married women who were highly educated (OR = 2.5, CI: 1.0 -6.1); who had financial capability (OR = 4.3, CI: 1.6 -11.3); who had medical insurance (OR = 4.2, CI: 1.4 -12.4); or who did not have previous birth experience (OR = 3.5, CI: 1.5 -8.5). Facility delivery was also more likely among married women who did not have the support of sisters-in-law for fetching water (OR = 2.2, CI: 1.0 -4.7), or who were advised to deliver at a health facility by family or neighbors (OR = 2.5, CI: 1.2 -5.5). --- Discussion The present study found that more than 80% of women delivered at health facilities, and less than 20% delivered at home. This is a high proportion of facility deliveries and low proportion of home deliveries in comparison to findings of the Kenya DHS (national estimate, 42.6% facility deliveries and 56.2% home deliveries) [9]. We believe this is because a large public health intervention focusing on maternal health had recently involved the current study site, thereby increasing the proportion of facility deliveries. Moreover, respondents were mothers who brought their babies to the health center for immunization. Thus we might have selected those who had already easy access to health facilities. However, a high proportion of facility deliveries is not unusual in urban areas within sub-Saharan Africa. Progress towards the professionalization of childbirth attendance in the urban population was already good in 1992 (>70% of births were attended by health professionals) and even better in 2000 [3]. A similar rise in facility deliveries would have accompanied this, since most professional assistance is provided in health facilities. On the other hand, rural sub-Saharan Africa showed no improvement; 32% of deliveries were attended by health professionals in the early 1990s and this proportion remained the same in 2000 [3]. While the present study site is not located in an area as urbanized as Nairobi, the capital of Kenya, some of its characteristics are more similar to urban areas than to rural areas because of economic development accompanying the tea industry. Bivariate analysis showed that place of delivery was associated with age, educational level, occupation, marital status, economic status, medical insurance cover, parity, support for daily tasks, and advice regarding facility delivery. Many studies have investigated factors influencing place of delivery. For instance, a study in Kenya [22] and one in Burkina Faso [23] reported that a high education level was associated with a high proportion of facility deliveries. Another study in African countries including Kenya showed that favorable economic status was a significant predictor of the proportion of facility deliveries [7]. Moreover, the present study found that women who had medical insurance were more likely to deliver at health facilities, in agreement with a previous study conducted in Kenya [24]. In addition, primiparae were significantly more likely than multiparae to deliver at health facilities, as in previous studies [25]. Women who have not previously given birth tend to be more worried about complications than those who have had previous deliveries; therefore, they tend to choose facility delivery. The current findings showed that types of support and the people providing this support were associated with place of delivery. Multivariate logistic regression analysis revealed that married women whose sisters-in-law helped them fetch water were less likely to deliver at health facilities. Bivariate analysis revealed that married women whose husbands supported them in farming and those whose neighbors helped them fetch water were also less likely to deliver at health facilities. These results are explained by the local cultural context. In the area of the study, women usually live near their husband's family after marriage, and mutual aid and collaboration with neighbors is quite common [26]. Having support from sisters-in-law, a husband or neighbors indicates stable relationships within the marriage, and with the husband's family and community. Our results showed that 44.8% of women who delivered at home were assisted by their mother-in-law (not TBA). Married women who had stable relationships with their in-laws readily received support to deliver at home, thereby decreasing their rate of facility delivery. Moreover, if a member of their husband's family recommends a home delivery to a married woman, it is difficult to act against such advice, since married women are under pressure from this side of the family. This supports the previously reported idea that social support can include negative support as well as positive support [27]. Some married women deliver at health facilities because of positive reasons such as believing it is safer than home delivery. A previous study found that the perceived competency of midwives and better equipment were among the reasons women used childbirth services [5]. On the other hand, some married women who do not have a good relationship with their husbands or their neighbors might deliver at health facilities for negative reasons such as not having support for home delivery. Hence, facility delivery may be chosen for both positive and negative reasons. The present study found that unmarried women whose mothers supported them in housework and whose sisters helped them fetch water were more likely to deliver at health facilities. Unmarried women are of course not subject to pressure from a husband's family, and they can readily get support from their original family, so they are less likely to be influenced by traditional custom related to home delivery. As a result, unmarried women who had social support were more likely to deliver at health facilities. The present multivariate analysis revealed that married women who were advised to deliver at a health facility by family or neighbors were more likely to do so than those who did not receive such advice. Bivariate analysis also showed that married women who were given such advice by their mother-in-law were more likely to deliver at a health facility than those who did not receive such advice. This suggests married women are more likely to be affected by advice from their mother-in-law than from any other family members. Mothers-in-law appear to be very influential regarding where their daughtersin-law deliver, and should therefore be involved with promotion of facility delivery. Moreover, married women who were advised by health staff to deliver at a health facility were more likely to do so than those who did not receive such advice, in agreement with a previous study [28]. However, the fact that women could receive advice from health staff suggests they already had good access to health services. Indeed, a previous study reported that professionals such as health care providers and counselors were not considered sources of social support by women [12]. Therefore, promotion of facility delivery should not accordingly be implemented only by health staff but also by close kinship. --- Limitations of the study Several limitations are worth noting. First, the respondents were mothers who brought their babies to the health center, raising the possibility of selection bias. Second, participants might have had recall bias about pregnancy and childbirth. Furthermore, women could have responded with socially desirable answers such as reporting that they considered facility delivery in a positive light. In addition, ascertaining support for housework, fetching water, and farming, and advice regarding facility delivery does not provide the whole picture of social support. We therefore measured only some of the instrumental and informational support available to women. Finally, we conducted multivariate analysis only for married women, because the number of unmarried women was inadequate. The generalizability of our findings may accordingly be limited. --- Conclusions In conclusion, our study found that social support was associated with place of delivery. This support took the form of instrumental support, such as support for daily tasks, and informational support, such as advice to deliver in a health facility. However, married and unmarried women differed in terms of the factors influencing their decision about where to deliver. Married women who had instrumental support were less likely to deliver at health facilities than those without this support. In contrast, unmarried women who had instrumental support were more likely to deliver at health facilities than those without this support. Married women who had informational support were also more likely to deliver at health facilities than those without such support. A woman's mother-in-law was the most influential person regarding place of delivery. We therefore need to recruit as advocates not only women of reproductive age but also those in the same generation as their mothers-in-law. These findings can be used by policy makers, planners, and health care professionals to take into account social support issues in improving health facility deliveries. Increasing the awareness of women and their family members about the benefits of facility deliveries are recommended. --- Included women who did not have any sisters and who were therefore not advised to have a facility delivery by their sisters. --- Additional files Additional file 1: Questionnaire1: questions about socio-demographic characteristics. Additional file 2: Questionnaire2: questions about birth experience. Additional file 3: Questionnaire 3: questions about social support. --- Competing interests The authors declare that they have no competing interests. --- Authors' contributions MO conceived of the study, did the data collection and analysis, wrote the paper. SH and AM also wrote the paper. SH and MK contributed to the study design. All authors read and approved the final manuscript.

Background: An estimated 358,000 maternal deaths still occur worldwide each year. The place of delivery is of great significance to the reduction of maternal mortality. Moreover, socio-economic factors, cultural traits, and local customs are associated with health-seeking behavior. This study aimed to explore determinants of association between social support and place of delivery.

Background Healthcare stakeholders have pronounced both enthusiasm and apprehension over the expanded use of real-world evidence (RWE) [1][2][3][4]. However, the patient communitythose who benefit from new treatments but are vulnerable to potential safety risks and whose routine medical interactions are used to generate RWE-has been less vocal [5][6][7][8]. RWE is information derived from studies analyzing realworld data (RWD), which refers to "those data derived from sources other than randomized controlled trials (for example, patient registries, patient cohorts, administrative claims, or electronic health records)." [4]. Emerging sources of data capturing the "real world" in terms of treatments, settings, and patient populations (e.g., subgroups) offer enormous opportunity for deeper understanding of why treatments work (or do not) and for whom [9]. Patients, as well as those making decisions on their behalf, such as regulators, value assessors, and clinicians, can benefit from research that better reflects patient experiences to support their own information needs [10][11][12]. RWE may help clarify the best uses for regulatory agency-approved treatments and add precision to value assessment and clinical decision making [13]. Despite these potential advantages, standards and quality varies across RWD sources and range of study designs [7,8,14]. Given patient ownership of their data and rights to privacy, it is also important to understand patient perspectives on RWE for ethical reasons [6]. Based on the premise that patients and their advocates may not be vocal or activated on RWE because of a lack of familiarity, in 2017, the National Health Council (NHC) and National Pharmaceutical Council convened a day-long multi-stakeholder roundtable discussion. The objective was to gather patient-community views on RWE, areas of greatest concern, policy issues, and information and tools most needed to understand, trust, and use RWE [5]. The purpose of this paper is to report insights from the discussion for consideration by patient groups, policy makers, clinicians, and researchers. --- Methods Invitations were extended to a convenience sample of NHC members primarily representing patient groups but also non-patient members with an interest in RWE (see the Electronic Supplementary Material [ESM] 1). Patient groups were oversampled to provide a majority of patient and patient-group representation, as these were the primary views sought. Representatives of non-patient stakeholder communities were invited because of their RWE content knowledge and to answer questions, provide examples, and support patient-participant engagement during each session to maximize the patient voice. Before the meeting, a briefing document with an overview of RWE definitions, key issues, and information from patient qualitative interviews was provided (see ESM 2). To open the meeting, participants heard context-setting remarks on RWE from a government regulator, health policy research center, large US payer with research capabilities, and a large patient organization. The opening panel was followed by three breakout sessions to gather patient-community perspectives on: 1. RWE, including but not limited to, issues of definition, concerns, transparency, privacy/security, sources, and meaningful use; 2. opportunities for improving the communication and dissemination of RWE to the patient community; and 3. skill sets and tools needed to understand and make the best use of RWE in decision making. Discussion questions are listed in ESM 3. Participants were organized into five small, pre-assigned groups, ensuring representativeness across stakeholders and patient representative leadership. In addition to a note-taker, an NHC staff member was assigned to each group as a facilitator and to assure that the patient voice was predominant. Participants, including note takers, were asked to prioritize patient perspectives, questions, and viewpoints in the discussion over other stakeholder views. Groups were asked to prepare storyboards summarizing key points, which were presented to the larger group at the end of each breakout session. Following the roundtable, NHC staff used storyboards, notes, and illustrative examples to develop a representative draft summary of results. The draft was circulated to participants for comment. --- Results --- Participants In total, 50 individuals participated, representing patient organizations (n = 21), industry (n = 12), research entities (n = 3), regulatory agencies (n = 1), and other healthcare organizations (n = 3). The resulting draft themes were circulated to 38 organizations, with six providing comment (two patient organizations, three biopharmaceutical organizations, one US government agency). Note that, in this context, patient refers to patient participants in our roundtable rather than patients at large, though typically patients and family caregivers are common among patient-group staff representatives. --- Patient Perspectives on Real-World Evidence (RWE) Our roundtable discussion identified ten themes with implications for policy makers, researchers, clinicians, and patient groups. --- Raising --- Enhancing Patient-Community Capacity --- Common Definitions for RWD and RWE are Vital Efforts to clearly define RWD, its sources, and how such data can be interpreted to yield useful RWE are needed. Patient advocates believe that better defining RWE is important to increasing their confidence in the data collection methods and applications. Education and tools to build capacity for patients to co-develop real-world studies or initiatives 10. Education and tools on shared decision making using RWE 11. Tools to help patient groups understand alignment of study rigor with the context of decision making 12. Tools for clinicians to use in interpreting and discussing RWE with patients like me"-as an assurance of how a treatment might benefit them personally. Studies should be designed in partnership with patients and with the intent to communicate results to patients in a balanced manner to inform decision making. Table 3 lists guiding questions that can improve trustworthiness and uptake of research findings. --- Skills and Tools are Needed by Patient Organizations to Facilitate Uptake of RWE --- Clinicians Must be Champions for Dissemination and Use Education efforts on RWE must target the clinician community, as clinicians are a primary conduit of information to patients. For RWE to be impactful, participants emphasized the need for clinicians to understand the value of RWE to identify potential treatment options and support patient decision making. --- Communications to Patients Should be Balanced and Empowering Roundtable participants identified important principles to guide communication of RWE to patients. These recommendations are summarized in Table 2. --- Contextual Considerations --- Acceptable Uses of RWE Must be Linked to the Context of its Use RWE may serve different purposes for different diseases (e.g., inform surrogate outcomes for future trials for rare conditions). Patient representatives generally agreed that RWE should not be used for clinical evaluation of new, unapproved therapies. However, patients saw opportunity for RWE to inform new uses of approved therapies in, for example, new patient subgroups (e.g., children), for treatment of comorbid conditions, and to achieve patientdefined endpoints not part of completed clinical studies (Table 3). --- Privacy Must be Protected and Data Ownership Clear to Promote Trust Patients often lack clarity about who owns data about their health and often struggle to access their own medical records. Patient groups were concerned about efforts that seek to commoditize data about them, without a clear focus on improving their health outcomes and care. While the Health Insurance Portability and Accountability Act (HIPAA) may allow for data sharing, privacy issues They can develop key questions and other "litmus tests" to assist with discerning the value of RWE for their constituencies. Patient-organization communication can be an effective method to build trust and convey information in language that is understandable and relevant to the audience Use varied communication methods tailored to the needs of the audience Such methods may include patient focus groups, peer-facilitated listserves or discussion groups, etc. Patient organizations can build on the narrative already at play in the community and find peer champions to convey messages and combat misinformation or junk science within the patient community. Collaborations within disease-state communities can ensure consistency and optimize resources Clinician/providers can be a resource regarding RWE While patient organizations are one important source about emerging evidence, clinicians provide another opportunity to convey such information and can help with translating RWE for their patients to understand and use in decision making. In turn, patients should seek their doctor's opinion about theories and information they are finding Patient-Community Perspectives on Real-World Evidence remain a concern for patients, and standards of protection to ensure anonymity must be clarified. --- RWE Should Include Authentic Sources of Patient Data Electronic health records and claims data, two primary sources of RWD, are only "real" to an extent and may not reflect the diversity of disease experiences, preferences, or outcomes. Participants recommended RWE should aggregate these data with additional real-world sources, including patient-generated data (e.g., patient-organization registries). --- Discussion Participants emphasized the importance of and opportunity inherent in the emergence of RWE. However, significant challenges are posed by study-design complexity and absence of clear standards for conducting and communicating RWE, especially to patients. These findings are largely consistent with research examining patient perspectives toward comparative effectiveness and patient-centered outcomes research. Law et al. [15] found that patients believe research should be (1) relevant to patient concerns, (2) presented in a way that is understandable to patients, and (3) useful for patient and clinician shared decision making. Actions to advance awareness and uptake of RWE are described in the following subsections. --- Investments in Capacity Building are Needed Individual patients look to patient groups for information to inform understanding of their disease and guide treatment decision making. Policy makers, advocacy groups, and other stakeholders need to invest time and resources into educational efforts. Similar to prior education initiatives to inform clinicians about emerging methods and data sources, patient-group-specific RWE training could help develop patient-community content knowledge, leading to more meaningful contributions to and uptake of RWE [16]. --- Patient-Friendly Summaries of All Research -Including RWE -Should be Standard Practice Researchers and policy makers can adopt policies to directly communicate with end users. For example, the Patient-Centered Outcomes Research Institute requires researchers to publish plain language summaries of their findings [17]. Wider adoption of patient-friendly summaries would help patient-group staff communicate RWE findings to their constituents. Summaries could be copied alongside links to the full articles in patient-group e-newsletters. --- Researchers Must Consider Patient Information Needs When Designing Studies To ensure that RWE is initially developed to inform patient and clinician questions, and that study designs reflect patient experiences to the extent possible, early patient-researcher partnerships are necessary. Researchers are encouraged to publish open access to avoid pay firewalls. Table 3 Questions patients want answered to aid in their understanding and trust of real-world evidence RWE real-world evidence --- RWE Should Complement Clinical Trial Evidence in Guidelines Patient understanding alone is necessary but not sufficient. The primary source of evidence for clinicians and patients is often via clinical practice guidelines. Standards for developing clinical practice guidelines recognize the limitations of focusing on only clinical trials, which may not reflect the diversity of patients or their experiences, and instead focus on high-quality evidence from both randomized trials and RWE [18]. Therefore, clinicians must also be educated on the interpretation of and the proper role for RWE in decision making [19]. --- Patient Registries can Supplement Traditional RWD Sources Unlike traditional RWD sources, patient-generated data sources include unfiltered information that has not been reinterpreted through clinicians [20]. These data sources may help researchers overcome limitations associated with electronic health record or administrative claims data [7]. For example, data on over-the-counter medications, frailty, or activities of daily living, which are frequently cited by researchers as important unmeasured variables, may be captured [19,21]. --- Limitations Participants were a convenience sample of NHC members, colleagues, and others who self-selected to participate. Patient representatives who participated were employees or lead volunteers of patient organizations and, while knowledgeable, were not necessarily patients or caregivers of patients with the disease of interest. This approach may not have captured the views of all stakeholders or the breadth of representation within the patient community. --- Conclusion Patient-system medical interactions have traditionally formed the basis for RWD. However, in the twenty-first century, passive patient engagement is insufficient. The themes we identified and the needs for moving forward will aid policy makers, researchers, clinicians, and patient groups in efforts to enhance understanding of, trust in, and engagement with RWE to inform decision making by all stakeholders. As researchers, policy makers, and regulators establish standards and processes for using RWE in regulatory, value, and clinical decision making, patients must be the primary focus for us to advance a patient-centered health ecosystem. --- Author Contributions JSG, JH, and EMP contributed to the conception and design of this work. EMO, JSG, JH, and EMP all contributed to the analysis and interpretation of data; data collection, drafting and revising the work; and final approval of the version to be published. All authors agree to be accountable for all aspects of the work. --- Conflict of interest

Background Healthcare stakeholders have pronounced both enthusiasm and apprehension over the expanding use of realworld evidence (RWE). The patient community-those who benefit from new treatments but are vulnerable to potential safety risks and whose routine medical encounters are used to generate RWE-has been less vocal. Understanding patient perspectives on the use of RWE to guide clinical decision making and inform regulatory decisions and value assessments is imperative. Methods We convened a day-long, multi-stakeholder roundtable in Washington D.C., USA, on 31 July 2017 to gather patient-community views on RWE and related concerns and the communications, information and tools needed by patients to understand, trust, and use RWE. Participants included a convenience sample of National Health Council (NHC) members primarily representing patient groups as well as non-patient members with an interest in RWE. Participants were organized into small, pre-assigned groups, ensuring representativeness across stakeholders and patient leadership. Discussions, including storyboards, notes, and illustrative examples were captured and later analyzed thematically by NHC staff. Results Ten RWE themes emerged: (1) most patients were unaware of RWE and its actual or potential uses, (2) common definitions for real-world data and RWE are needed, (3) patient organizations need RWE skills and tools, (4) patient-scientist partnerships can help differentiate high-quality RWE, (5) RWE should inform decision making, (6) clinician support is needed for RWE uptake in patient decision making, (7) communications to patients should be balanced and empowering, (8) context of use impacts RWE acceptability/trust, (9) privacy/data ownership require clarity, and (10) patient-generated data are also real-world data (RWD). Conclusion Patients see great possibility in using RWE to understand how a treatment works-to find someone that "looks like me" as assurance of how a treatment might benefit them personally. Patient groups will play a critical role in helping to educate constituents on understanding, contributing to, and using RWE. To maximize patient uptake and the co-development and application of RWE, patient groups require education and tools.

Part C of the Individuals with Disabilities Education Act (IDEA) authorizes states to provide early intervention (EI) for infants and toddlers with or at risk for developmental delays and disabilities with the incentive of federal financial support. 1 EI is not a single program, but rather includes different packages of services that vary within and across states with regard to delivery models and provider disciplines. 2 This federal-state partnership allows flexibility to address the diverse medical, developmental, and social needs of families. State flexibility in EI resource allocation and programming has the potential to ensure appropriate EI referrals and to reduce EI access and service use disparities. In this paper, we describe social disparities in EI service use. Social disparities refer to differences in health and healthcare access that are systematically distributed and stem from varying levels of social disadvantage. 3 For example, despite state flexibility in EI spending, previous literature suggests EI access disparities based upon social and developmental risk. 4,5,6,7 The mechanisms through which families enroll in EI are multi-faceted and in addition to the diagnosis of a developmental condition, also typically require appropriate community outreach, good communication from pediatric providers, and optimal parental knowledge and supportive beliefs about developmental and therapeutic intervention. This constellation of factors influencing EI enrollment can contribute to access disparities. Indeed, a number of studies have shown that children who are black, non-Hispanic, 4 have a developmental delay, 5 and those who are poor or publically insured 3,5,6 are less likely to access EI than children who are white, non-Hispanic and from families with less social risk. Ensuring access to coordinated systems of care for young children with developmental conditions is a national priority for reducing service disparities. For example, the American Academy of Pediatrics endorses universal developmental screening to help facilitate timely EI referral for infants and toddlers with developmental conditions and suspected delay. 8 Moreover, many states have adopted Maternal and Child Health Bureau performance measures 9 to ensure appropriate access to EI for children with developmental conditions and delays. Disparities in EI access may contribute to inequities in EI service use and outcomes, yet there is limited research evidence regarding optimal EI dosage. Although EI programming varies state to state, core EI services exist in all states and include physical, occupational, and speech therapy, and special instruction (i.e., early childhood special education or developmental specialist) while some programs also provide psychology services. Previous research from a national EI sample 10 provides descriptive evidence about EI service use. Specifically, about one-third of children receive up to one hour of EI services per week, nearly one-third of children receive 1-2 hours of services per week, and one-third receive more than two hours of services weekly. Moreover, 40-50% receive therapy (i.e., physical, occupational, or speech therapy) EI services. However, these data did not examine service use differences by child and family characteristics. More recently, analyses of data from a state early intervention program 11 suggested that children who received greater number of hours and intensity of EI services were more likely to be younger at EI entry, insured, and non-poor. However, this study did not examine differences in EI use by type of core EI service, and did not examine differences in EI service use by racial and ethnic sub-groups or developmental condition. There is evidence to suggest that minority, poor, under insured children and those with more severe conditions were least likely to have their therapy needs met. 12,13 However, these studies examined unmet need for therapy, generally, and was not specific to EI so it is unclear if similar disparities would persist among an EI-enrolled cohort. Finally, there is a paucity of extant literature linking EI service use with outcomes. Prior research has described associations between child and family characteristics and parent-reported outcomes 14 and improvement following EI services overall. 9 Yet, it is not clear how service use (i.e., type, amount, duration, or intensity) is associated with more optimal EI outcomes. Greater specificity in knowledge about disparities in EI service use and outcomes may highlight opportunities to customize and coordinate care for improved EI care quality. This study was undertaken in order to leverage existing electronic administrative data on discharged families in order to describe specific EI service use patterns and outcomes. The primary purpose of this study is to describe disparities in discipline-specific EI service use patterns. We also sought to describe the association between EI service use (overall and discipline-specific) and provider perceptions of the child's improvement upon EI discharge. --- METHODS --- Study Sample The study sample was derived from electronic administrative data at an EI program in XXX, Colorado. This EI program is located within the XXXX (Blinded). The EI program has a 20year history of providing EI services to approximately 125 families annually. The study sample included children who met the following inclusion criteria: 1) discharged from the EI program between 7/1/2010 and 6/30/15, and 2) had at least one EI service. --- Measures EI Service Use-The primary outcome of interest was EI service use. In this study, EI services most often occur in the families' homes. Although the services are provided during typical business hours, efforts are made to accommodate family schedules (e.g., evening home visits). Overall EI service use was described according to number of services (sum of discipline-specific services received [1, 2, 3 or more]), total number of EI hours, and total duration of EI services (in months). Overall EI intensity, categorized as the number of all EI service hours per month, was calculated as the total EI hours divided by the total duration of EI services (in months). To estimate discipline-specific service use (i.e., the type of service, according to the provider discipline, for example physical therapy), we estimated 1) whether children received "any" discipline-specific services and 2) the intensity of each disciplinespecific service. Children were categorized as receiving "any" discipline-specific service if they received at least one billable service from that discipline (e.g., physical therapy [PT], occupational therapy [OT], speech therapy [ST], psychology, or developmental specialist [DS]. A DS typically has a background in early childhood special education or early childhood development. Discipline-specific service intensity was calculated as the total number of service hours for each discipline, divided by the total duration of EI services for that discipline (in months). Provider Rating of Child's EI Improvement-A secondary outcome of interest was providers' rating of the child's EI improvement. This single item is routinely administered at the time when the child's EI record is archived and the Child Outcomes Summary form 15 and exit developmental evaluation have been completed. The single-item rating captures EI provider perceptions of the child's overall progress between EI entry and EI exit (5-point scale, ranging from 1 [no improvement] to 5 [excellent improvement]). This item was developed by the EI program director and intended for internal use. The item has been routinely administered for nearly 10 years, and all program staff received training from the program director. There is no reliability and validity evidence for this measure, but sample score distribution is similar to national estimates of children's progress between early intervention and exit. 16 Child and Family Characteristics-Reason for the child's EI eligibility was categorized according to child's developmental condition type. Specifically, children are eligible for EI based upon a developmental condition, (i.e., an established condition or developmental delay). An established condition is a diagnosis that typically results in developmental delay so the child is automatically deemed to be EI eligible. These conditions are based upon federal legislation 1 and states can choose to include additional diagnoses. The EI program engaged in this study draws on their state's database to identify those developmental conditions that are associated with automatic EI eligibility. 17 Moreover, in CO, children are eligible for EI based upon the presence of a developmental delay, determined by a multidisciplinary developmental evaluation administered by the EI team. Child's race and ethnicity was categorized as white, non-Hispanic (WNH); black, non-Hispanic (BNH); Hispanic; and other race, non-Hispanic (ONH; includes Asian, Pacific Island and children reporting more than one race). Highest parental education level was grouped as 12 years or less; between 13 and 15 years; and 16 or more years. Annual household income was categorized as less than $50,000 or $50,000 or more. We also included a measure of child's sex (male/female) and the primary language spoken at home (English vs a language other than English). Insurance type was categorized as private or public (includes Medicaid and Children's Health Insurance Program [CHP+]). --- Analytic Approach We first calculated descriptive statistics for select child and family characteristics and EI service use. For continuous variables, we calculated sample means, medians, and standard deviations. For continuous variables with a non-normal distribution, we additionally calculate inter-quartile ranges. Sample proportions were calculated for each categorical variable. We described the associations between child and family characteristics and EI service use in two ways. First, we estimated unadjusted odds of receipt of "any" of discipline-specific EI services, controlling for each child and family characteristic, by estimating a series of bivariable logistic regression for each discipline. To estimate the unadjusted associations between each child and family characteristic and EI intensity, we fit a series of bi-variable median regression models. We first fit a series of models describing overall EI use and then fit a series of models describing service use by professional discipline (i.e., PT, OT, ST, and DS). Given the small sub-sample of children who received services from a psychologist, a model examining predictors of any (yes/no) or intensity of psychology services was not estimated. Median regression models the median rather than the mean and is appropriate for skewed data. We then estimated, for each discipline-specific service, a multivariable logistic regression model that included all child and family characteristics. These models represent the adjusted odds of receiving "any" of the discipline-specific services. Finally, we estimated a series of multivariable median regression models, first describing overall EI intensity and then fit a series of models describing EI intensity by discipline. In order to describe the association between service use and provider-reported global improvement, we estimated a series of ordinal logistic regression models. Ordinal logistic regression models are appropriate for ordinal data and the model coefficients are interpreted as the association between each child and family characteristic on the odds of demonstrating a higher (i.e. more optimal) improvement category than a lower one (i.e., less improvement). The first model included EI intensity in addition to the child and family characteristics. Three additional adjusted ordinal regression models included discipline-specific (i.e., PT, OT, and ST) EI intensity as well as the child and family characteristics. These disciplinespecific models were conditional on receipt of "any" of the respective discipline-specific service. Given the small sub-sample of children receiving DS, a separate model examining the association between DS intensity and provider-reported global improvement was not estimated. The final adjusted ordinal logistic regression model included count of EI providers, EI intensity, and the child and family characteristics. For the adjusted ordinal regression models, we present an odds ratio and 95% CI for each model parameter. --- RESULTS The sample included 457 children as described in Table 1. Children were predominantly male (65.7%), white, non-Hispanic (59.7%), and spoke English as their primary language (89.5%). More than half of the parents had at least 16 years of education (54.1%) and earned $50,000 or more annually (58.3%). Nearly 1/5 (17.7%) of the sample was publically insured, and about 1 in 5 (22.1%) children had an established condition. [10.4, 45.9] for DS, 7.5 [5.5, 16.1] for psychology, 24.5 [14, 40] for ST, 18.5 [8, 34] for OT, and 35 [16.5, 62] Results of the adjusted logistic regression models estimating the odds of receipt of each discipline-specific EI service are presented in Table 2. As compared to children with a developmental delay, those with an established condition have between 2-and 3-fold higher odds of receiving physical therapy (OR=2.3 95% CI=[1.4, 3.9]), occupational therapy (OR=2.5 95% CI= [1.5, 4.3]), and special instruction services (OR=3.5 95% CI=[1.6, 8.0]). Children whose annual family income is at least $50,000 have twice the odds (OR=2.0, 95% CI=[1.0, 3.9]) of receiving physical therapy as compared to their peers in the lower income category. Children whose primary language is English have lower odds of receiving PT (OR=0. The results of the adjusted median regression models estimating the association between child and family characteristics and intensity of EI services are presented in Table 3. Compared to children with a developmental delay, those with an established condition receive nearly an hour more per month of EI services (b=0.8, SE=0.2). Compared to children whose parents have at least 16 years of education, children whose parents have 12 years of education or less receive approximately 40 more minutes of EI services per month (b=0.6, SE=0.3). The results of the adjusted median regression models estimating the association between child and family characteristics and intensity of discipline-specific EI services are presented in Table 3. Compared to their WNH counterparts, BNH children receive about 45 fewer minutes (b=-0.7, SE=0.4) of PT per month. Compared to children with a developmental delay, those with an established condition receive a higher intensity of PT (b=0.6, SE=0.2) but a lower intensity of ST (b=-0.7, SE=0.4). Compared to children who are privately insured, children with public insurance receive a lower intensity of OT (b=-0.5, SE=0.3) and ST (b=-0.6, SE=0.3). Compared to their WNH peers, Hispanic children receive nearly an hour per month more of OT services (b=1.0, SE=0.3). Finally, compared to children in the lower income bracket, children whose family income is at least $50,000 per year receive nearly one more hour of OT per month (b=0.7, SE=0.3). --- As shown in The results of the adjusted ordinal regression models estimating the association between EI service use and provider-reported improvement at EI discharge are presented in Table 4. Children who are ONH had a nearly 70% lower odds of having higher global improvement (OR=0.4, 95% CI: 0.2, 0.8) even controlling for EI intensity child and family characteristics. In addition, in the discipline-specific models children of an "other" race who received OT (OR=0.3, 95% CI = [0.1, 0.8]) each had nearly 70% lower odds of receiving a better global improvement score than their white, non-Hispanic counterparts receiving OT. For children who are of an "other" race, these results held even after controlling for child and family characteristics, overall EI intensity, as well as provider count. --- DISCUSSION This study used a single administrative database to describe EI service use patterns among children who had received these services from a single program located in XXX (Blinded), Colorado. Results of this study suggest that sample children on average receive about 2 hours of EI services per month, however there was substantial variability in the amount of EI services received. The most prevalent core service type was speech therapy but the core EI service with greatest total hours was physical therapy. These results are similar to national estimates 9 suggesting that over half of EI-enrolled children receive speech therapy. Moreover, our findings are consistent with the variety of service intensity received in EI where, among a nationally representative sample of EI children, about 1/3 receive less than 1 hour per month; another 1/3 receive 1-2 hours per month, and 1/3 receive greater than 2 hours per month. 9 Results of this study also suggest significant social disparities in EI service use generally and by most EI disciplines. For example, compared to their white, non-Hispanic peers, BNH children, most with developmental delay, were not only less likely to receive any PT, but when they did receive PT, it was of a lower intensity. Our study findings are consistent with previous research suggesting that black, non-Hispanic children, especially those with a developmental delay (i.e., versus established condition) are significantly less likely to access EI than their white, non-Hispanic counterparts even controlling for condition severity. 4 Moreover, previous research also suggests parents of minority children with developmental conditions report a nearly 2-fold higher unmet need for therapy compared to similar white, non-Hispanic parents. 19 To our knowledge, ours is the first study to describe disparities in EI service use generally and with respect to discipline-specific service use. While the reasons for these differences in service use could not be ascertained from this study, results are similar to those reported in prior studies related to EI service access. A recent review by Zuckerman and colleagues 18 examined the extant literature to describe disparities in receipt of early childhood services for children with developmental conditions. Those results suggest that racial differences in care stem from differences in provider behavior and parental perceptions of child development. For example, minority children are less likely to have timely access to primary and specialty care. Limited health care access could translate into fewer opportunities for parents to express developmental concerns and a lack of provider follow-up about developmental concerns based on an abnormal developmental screen that would result in EI referral. Moreover, disparities in healthcare access could mean that minority families are less likely to have a medical home 19 and receive information about the importance of EI therapy participation. As such, these missed anticipatory guidance opportunities might translate to BNH children receiving less physical therapy than their WNH peers. Zuckerman and colleagues also suggest that healthcare providers are more likely to dismiss parental developmental concerns and abnormal developmental screening results to social risk (i.e., versus medical or developmental condition) for BNH families than WNH, resulting in fewer EI referrals and follow-up regarding developmental services. These provider behaviors could undermine EI outreach efforts to BNH families who are eligible for EI but not identified this way as often in the context of a primary care visit. Finally, cultural differences in parental expectations or their concerns about the face validity of developmental screens that have been critiqued for inherent racial biases, [18][19][20][21] may decrease how often parental concerns are reported by BNH families. Interestingly, we found that Hispanic children received greater OT intensity than WNH peers. Follow-up conversations with the EI program suggest that many staff are bilingual and able to communicate with families in Spanish, perhaps mitigating an important therapy service use barrier for Hispanic families. Future research should explore barriers to EI physical therapy use for BNH families and optimize EI participation. Moreover, we found that children from families with higher incomes are more likely to receive physical therapy and receive higher intensities of occupational therapy than their low-income counterparts. It is plausible that families with higher incomes have access to additional resources to have greater knowledge, beliefs and expectations related to therapy service use and intensity, which allows them to advocate for greater therapy service delivery. Additionally, families with higher incomes likely have additional resources (e.g., flexible employment, paid time off etc.) that allow for better adherence to a therapy plan of care, particularly services that most often occur during typical work hours and place high demands on parent involvement. For example, most EI therapeutic interventions use mostly face-to-face methods for parental engagement, which requires that the parent be physically present during the EI session. We also found that children who are publically insured receive less intensive speech therapy than their privately insured counterparts. This finding could be explained, in part, by the finding that children with a developmental diagnosis are less likely to receive speech therapy. That is, public insurance might be marker for not only social disadvantage, but also developmental disability (i.e., children who are dually eligible for social security disability insurance and Medicaid, for example). To this end, children with developmental delays may be accessing speech therapy more often and at greater intensities than their counterparts with established conditions. Indeed, in previous research examining EI relevant sub-groups, 20 children with mild developmental delays consistently appear as a clinically and policyrelevant group with nearly 30% of EI-enrolled children having a speech delay and receiving primarily speech therapy services. Further research should explore the extent to which children with developmental disabilities, particularly those who are publically insured, have their speech therapy needs met. Yet, receipt of public insurance could also be a marker for relative social disadvantage, particularly given that health insurance is not required to access to early intervention and does not drive how much therapy a child receives. Thus, publically insured and low-income children may lack availability of family resources to adhere to a more intensive occupational, speech, or physical therapy program, In this study, children with developmental condition that automatically qualifies them for EI (i.e., established condition) receive greater EI service intensity overall, more intensive physical therapy services, but less intensive speech therapy. This finding is consistent with previous literature 5 suggesting that children with an established condition access EI more often. However, these findings can also contribute to the extant literature to describe service use (i.e., intensity) differences by condition type. However, previous literature also suggests that parents of children with established conditions report more unmet need for therapy than parents of children with developmental delays. Thus, future research should link EI service needs, service use patterns and outcomes to better understand and mitigate disparities in unmet need for children with varying developmental conditions. In this study, we did not find that service use was associated with perception of improvement. However, we did find that greater intensity of ST services was associated with small odds of lower provider-reported improvement. The positive relationship between service use and outcomes has been reported previously, 9 but did not fully adjust for child and family characteristics (e.g., family income, maternal education) that confound this association. Moreover, previous research did not address intensity, which is likely a better metric of service use than amount or duration of EI services. However, we acknowledge the limitations of a single item provider-reported outcome measure. It is possible that biases are introduced when a provider who has been working with a family for an extended time is more likely to report improvement at EI discharge. Alternatively, the provider assigned to this task varies and providers who are less familiar with a child may be less informed and therefore more likely to report improvement. Moreover, differences in provider reported outcomes may also be influenced by provider's expectations for improvement based upon the family's resources. Although the possibility of bias exists, it is minimized because this scale is administered for general program surveillance (i.e., not reported externally) and is reported annually in aggregate by the data manager. Thus, specific scores are not shared with families nor is a specific score tracked to a specific provider-child dyad. An additional limitation is that this measure was collected by provider report and may not capture parental perceptions of the child's overall progress. Future research should explore the role of collecting EI outcome measures that have particular relevance to families. Related, future research should examine the extent to which EI general and discipline-specific service use is linked to attainment of goals delineated in the child's EI care plan (i.e., Individualized Family Service Plan). We acknowledge the study's limitations. The data were ascertained from one early intervention program, which limits the generalizability of the findings to other EI programs within or outside of Colorado. Yet, where available for comparison, our sample resembles state and national estimates, and our main findings (e.g., percent receiving core disciplinespecific services) are consistent with national estimates, 9 which increases their external validity. Future research should expand the outcome research methods in this study to examine statewide and multi-state trends in EI service use disparities. Related, it is possible that EI service delivery models (e.g., multidisciplinary, transdisciplinary) influence EI service use data. Per national recommendations, 21 this EI program utilizes a transdisciplinary model and future research should investigate the role of service delivery models on EI service use and outcomes. Secondly, this study identifies social disparities in service amount but not EI care quality according to best practices. 21 Differences in EI service processes (e.g., family-centered care, care coordination and service delivery model, cultural competence) and family engagement will be important to investigate in future research examining social disparities in EI service use, results of which may guide best practices in early intervention. Finally, we included only one, single-item measure of children's global improvement. As noted, this scale has not undergone psychometric testing and may have innate biases. Yet, the purpose of this study was to leverage available data and understand opportunities for improvement in outcomes measurement. Future research should examine the feasibility of implementing additional measures and perspectives on improvement that could capture specific areas of gain. This study has a number of strengths. Access to a rich EI program database allowed us to describe a broad range of EI service use metrics. We were able to describe service use intensity overall and by service type and our analyses were able to control for a host of child and family characteristics. To this end, the study's findings highlight important racial and ethnic disparities in service use. Finally, the methods used in this study can serve as a model that could replicated by other EI programs individually or at the state-level to investigate and mitigate service use disparities. --- CONCLUSION The results of this study highlight substantial disparities whereby BNH, low-income, and publically insured children receive less intensive EI therapy services that their WNH peers with less social risk. The findings have important clinical and policy implications for understanding how to best tailor EI messaging and service delivery to assist families in navigating the EI system most efficiently and effectively with an eye toward reducing disparities in care and outcomes. For example, EI providers may need to employ more intensive or alternative strategies to connect with poor and BNH families and to ensure that EI services are well-matched to the family's needs, priorities, and desired outcomes for their child. --- Author Manuscript Khetani et al. --- Author Manuscript Independent Variables Adjusted Associations [0.

Objective-To describe social disparities in early intervention service use and provider-reported outcomes. Methods-Secondary data analysis of administrative data to ascertain EI service use of all EI and discipline-specific services and child and family characteristics. Adjusted logistic regression models estimated the odds of receiving each type of core EI service. Adjusted median regression models estimated differences in EI intensity for each type of core EI service. Adjusted ordinal regression models estimated the association between each type of EI therapy service and provider estimates of children's global functional improvement. Results-Children with a diagnosis (b=0.8 SE=0.2) and those whose caregiver had 12 years of education or less (b=0.6 SE=0.3) had higher EI intensity. Black, non-Hispanic (BNH) children had nearly 75% lower odds of receiving physical therapy (PT) [OR=0.3, 95%CI:0.1, 0.7] and greater odds of receiving speech therapy [OR=3.4, 95%CI:1.3, 9.3] than their white, non-Hispanic (WNH) peers. BNH who received PT received about an hour less per month (b=-0.7 SE=0.4) than their WNH peers. Hispanic children [b=1.0 SE=0.3)] and those with higher family income (b=0.7 SE=0.3) received greater intensity of PT compared to their peers who are WNH and from lowincome families. Publically insured children had lower intensity of OT (b=-0.5 SE=0.3) and ST (b=-0.6 SE=0.3). Greater intensity of EI services was not associated with greater provider perceived improvement.suggest disparities, by race and family income, in receipt of EI therapy services. These findings highlight opportunities to customize and coordinate care for improved EI access and care quality.

Background The Federation of National Organizations Working with the Homeless (FEANTSA) [1] defines homeless persons (HPs) as those unable to achieve or maintain adequate and permanent accommodation adapted to their situation, either due to economic reasons, social barriers, or the inability to lead an autonomous life. In Spain, the National Statistics Institute (INE) [2] reported that around 23,000 HPs visited social support centers in 2012, although it is estimated that more than 30,000 HPs in the country were not included in these figures [3]. In 2020, a mean of 17.772 HPs/day were accommodated by social services, 1.3% fewer than in 2018 [4]. In the study by Fazel, Geddes, and Kushel [5] on the health of HPs in high-income countries, homelessness was related to poverty, family problems, mental health issues, substance abuse, and/or structural factors such as the lack of low-cost homes. Little research has been conducted on more subjective aspects such as the general happiness or wellbeing of HPs, which may be useful to support the design of more effective interventions [6]. Ahuja et al. [7] found that their subjective well-being was inferior to that of people with homes. The health conditions of HPs pose a major care challenge to nurses working in the community, and an approach that accounts for HPs' own perception of wellbeing and happiness may improve the effectiveness of interventions in this population, such as those that focus on happiness [8,9]. Obtaining information about and analyzing happiness in general is a highly complex task. Easterlin [10] considered that subjective indicators are useful to evaluate happiness, but the optimal approach remains under debate. A study of subjective well-being in 11 nations found that it was correlated with their social, economic, and cultural characteristics [11]. Layard [12] studied happiness in the general populations of 50 countries and concluded that it was influenced by seven main factors: family relationships, economic status, work, community, friendships, personal freedom, and personal values. All except for the economic factor refer to relationships. The quality and stability of relationships and concern for others have been reported to generate greater happiness in individuals than concern for themselves [13], leading to greater success in life [14], One characteristic of HPs is that they vary in almost all the specific vital controls. In a study of 235 HPs in Madrid (Spain), 46.8% reported feeling happy, and a general state of happiness was associated with: not feeling alone or abandoned, not suffering disability or severe or chronic disease, having good expectations for the future, identifying with some religious belief, and having a positive perception of their own health status [15]. However, a study of 20 HPs in Australia found that health contributed little to their general perception of subjective wellbeing, which was more closely associated with feeling safe, being positive, feeling good, connecting with others, and participating in "normal" life [16]. A study of the characteristics of HPs in Granada (southern Spain), conducted by the present researchers in 2017, found that 6 of the 15 HPs then interviewed had been living on the street for more than eight years, and that 9 of them felt happy, a surprising finding given their difficult life conditions [17]. These observations, and the contradictory findings in the literature, prompted the development of the present investigation in the same city, using questionnaires and interviews to gain knowledge of the most important concerns of HPs, the reasons for their happiness or unhappiness, and their expectations for the future and satisfaction with their life, as expressed in their own words. The aim was to improve the relationship between HPs and care professionals, enhancing the trust needed for effective health care interventions. The present study was comprised of two phases. The objective of the first phase was to determine the characteristics of HPs in Granada and to analyze the relationship between these characteristics and their self-perceived happiness. The objective of the second phase was to explore the factors that HPs considered to affect their happiness or unhappiness. --- Methods --- Design In the first phase, an observational analytical study was conducted to determine the characteristics of HPs, their happiness status, and other study variables, using questionnaires. In the second phase, a qualitative descriptive study was carried out to explore the experiences of HPs using a content analysis of semi-structured interviews. The study was conducted between April 2017 and February 2018 in Granada (Spain). --- Sampling The enrolment of participants started after a two-year period (2016-2018) in which the principal researcher and other concerned individuals made regular nocturnal visits to places where people were sleeping on the street in the city of Granada, making contact with the homeless. Workers in the Asociación Calor y Café (Heat and Coffee Association) participated in these visits and became involved in the present project. This association runs a center used by many HPs in Granada that provides breakfast and an afternoon snack, facilities to wash and clean clothes, and leisure activities, as well as social support. Before the start of the present study, the researchers spent a large amount of time at this center to gain the trust and confidence of the users. Participants in this study were HPs who visited the Calor y Café center or were found in the streets of the city. An intentional consecutive non-randomized sampling procedure was conducted that took account of no other characteristics, selecting 25 HPs who attended the Calor y Café center and consented to study participation and the audio recording of interviews. Exclusion criteria were signs of drunkenness and/or mental confusion at the time of the interview. A nocturnal survey of HPs in Granada in 2019 found 62 people sleeping on the street, without counting those sleeping in caves or abandoned buildings; the majority were aged between 46 and 50 years, and 89.3% were men [18]. --- Data Collection The data were collected by two researchers working separately, who approached HPs in the street or while they were in the Calor y Café center for breakfast or an afternoon snack. After receiving an explanation of the purpose of the study and its voluntary character, their consent was sought for the completion of questionnaires and an audio-recorded interview. They were informed that all data would be treated anonymously in accordance with national and European data protection legislation (Spanish Law 15/1999, 13 December; EU regulation 2016/679, 27 April 2016). In the first phase of the study, the questionnaire included the sociodemographic variables: age (years), years on the street, number of children, sex, origin (Andalusia, rest of Spain, foreigners), educational status (no studies/primary school, secondary school, professional training, university degree), causes of homelessness (economic crisis, divorce, family conflicts, other causes), health (good health, poor health), whether participant had been assaulted (yes, no), where the participant sleeps (on the street, in temporary lodgings), source of income (no income, non-contributory pension, temporary employment), jobseeker (yes, no), requested assistance from social services (yes, no), and whether he/she feels discriminated against (yes, no) and feels happy (yes, no). After an interval of from two to three months, the above participants were contacted, when possible, for the second phase of the study, which was conducted at the Calor y Café center by the same researchers. The following questionnaires were administered: -Diener's Satisfaction with Life Scale (SWLS): This instrument has been validated in Spain [19] del instrument original [20]. It includes five questions on the responder's overall satisfaction with life, with Likert-type responses ranging from 1 (completely disagree) to 5 (completely agree). -General Happiness Scale: Visual scale evaluating general happiness in seven categories, represented by pictures of expressive faces: very happy, rather happy, somewhat happy, neither happy nor unhappy, somewhat unhappy, rather unhappy, and very unhappy. Participants respond to the question: Which of the following faces best represents your level of general happiness? This scale was used in the study by Vazquez et al. [21] -Goldberg's Anxiety and Depression Scale [22], using the version adapted to Spanish by Montón, Pérez Echeverr<unk>a, Campos, Garc<unk>a Campayo, and Lobo [23]. This has two subscales, one for anxiety and the other for depression, classifying responders as having or not having "probable anxiety" (cutoff: <unk>4) and/or "probable depression" (cutoff: <unk>2). -A further question was added regarding their expectations for the future (improve/remain the same/worsen). After completing the above questionnaires, a semi-structured interview was conducted, audio-recorded and transcribed. All participants were asked "What is happiness for you?"; if the response was positive, they were asked: "What makes you happy?", and if negative, they were asked "What would you need to feel happy?". Next, they were asked "What time of your life do you remember as happiest?" and, finally, "Do you feel good about yourself?". Interviews were individual, private, face-to-face, and of variable duration according to the situation. The researchers were previously trained to ensure an unprejudiced neutral attitude towards the homeless. The digital recording of each interview was downloaded in a password-protected computer and transcribed by the same researcher. The study complied with EU regulations (2016/679) and Spanish legislation (3/2018) on personal data protection and digital rights and was conducted in accordance with the 2013 revision of the Declaration of Helsinki (https://www.wma.net/what-we-do/medicalethics/declaration-of-helsinki/; accessed on 13 July 2017). All subjects gave their informed consent to participate in the study, which was approved by the clinical research ethics committees of Andalusia on 30 May 2017. --- Data Analysis A descriptive analysis was conducted, calculating the frequencies and percentages for qualitative variables and medians [P25-P75] for quantitative variables. The origin of HPs, their educational level, their relationship with partner and/or offspring, cause of homelessness, source of income, satisfaction with life, and happiness results were all treated as dichotomous variables to permit bivariate analyses, given the limited sample size. Fisher's test was used to evaluate the relationship of happiness results with qualitative variables and the non-parametric Mann-Whitney test for their relationship with quantitative variables. Spearman's rho test was applied to determine the relationships among quantitative variables. A p <unk> 0.05 was considered significant. IBM Corp. Released 2010. IBM SPSS Statistics for Windows, Version 19.0. Armonk, NY, USA: IBM Corp was used for data analyses. Qualitative analysis was conducted by a panel of three researchers following a descriptive content analysis procedure [24]. After reading the transcribed interviews, they encoded the most frequent responses and their number, organizing them into descriptive categories as a function of the main themes. Finally, the reliability of the design, data gathering, and data analysis was examined using a triangulation procedure with two external examiners for the encoding and identification of categories and themes [25,26]. The project was approved by the Ethics Committee of the Autonomous Community of Andalusia. --- Results Of the 30 HPs contacted for the initial quantitative phase of the study, 3 were excluded due to their condition at the scheduled time of the interview and 2 refused participation, leaving a final sample of 25 HPs. Of these 25 HPs, it was possible to make contact with 15 for the second phase of the study, which had a final sample of 14 HPs after excluding one participant for signs of mental confusion at the interview. The median age of participants was 54 years, and 20 were males. They had lived on the street for a median of 0.9 [0.2-8.5] years, 13 (52.0%) had secondary or higher education, 20 (80.0%) had children, 9 (36.6%) attributed their homelessness to the economic crisis, 12 (48.0%) reported poor health, 15 (60.0%) had been assaulted on the street, 14 (56.0%) received no economic income, 12 (48.0%) were not seeking employment, 18 (72.0%) had not asked social services for assistance, 14 (56.0%) did not feel socially discriminated against, and 16 (64.0%) felt happy (Table 1). There were no significant differences in any study variable between those who felt happy and those who felt unhappy (Table 2). In the second phase of the study (n = 14), which included twelve men and two women, ten participants were classified with probable anxiety and twelve with probable depression; five reported being satisfied with life (grouping together the categories highly satisfied, rather, and somewhat satisfied); seven reported being happy (grouping together Happiness Scale categories very happy and rather happy"; and eleven had future expectations of improvement (Table 3). A significant association was found between life satisfaction and happiness (p = 0.021). HPs who were satisfied with life were 4.5-fold more likely to feel happy than those who were not (OR 4.5). There was a close-to-significant tendency of higher anxiety among those who felt unhappy in comparison to those who felt happy (100.0% vs. 42.9%, p = 0.07). No association was found between expectations and happiness (p = 1) or between expectations and satisfaction with life (p = 0.72) (Table 4). The sole significant correlation between quantitative variables was a negative correlation between the number of arrests and the satisfaction with life score, i.e., the more arrests, the lesser the satisfaction (rho = -0.655, p <unk> 0.05). --- Analysis of Interviews The analysis of interviews yielded three main themes that organized the discourses on happiness: conditions for happiness, their own happiness/unhappiness, and self-esteem (Table 5). 1. Theme: Conditions for happiness We identified three categories that organized the conditions for happiness as indicated by participants: material, affective, and subjective (Table 5). Material conditions were described in terms of "having" objects, states, or situations, and were indicated by four participants. Affective conditions were emotions and affections related to "love" in the sense of giving, receiving, and sharing, indicated by four participants. Subjective conditions were personal conditions, attitudes, beliefs, and/or moods, indicated by eight participants (citations in Table 5). --- Theme: Own happiness/unhappiness This theme was related to different aspects related to their happiness/unhappiness articulated around "feeling of current happiness" and "feeling of unhappiness". The former was expressed as their feelings at the time of the interview and what made them feel happy; six participants reported feeling happy and two partially happy. The reasons for feeling happy were mainly affective and subjective. The reasons for feeling unhappy were mainly related to subjective, emotional, and health aspects (citations in Table 5) 3. Theme: Self-esteem This theme was related to self-satisfaction as an expression of inner coherence that can be part of the feeling of happiness. Discourses on this subject were generally brief and monosyllabic (in seven participants), contrasting with the more fulsome explanations on other issues. Responses to this theme were positive from nine participants, negative from three, and ambiguous from one (citations in Table 5). --- Discussion This study of HPs found an association between satisfaction with life and happiness, in agreement with various studies in the general population [27,28]. Satisfaction with life is a cognitive evaluation of the quality of one's own experiences, an indicator of subjective wellbeing expressed by the individual [29]. The main components of subjective wellbeing are considered to be emotional responses or affections and satisfaction with life [21]. In their systemic review, Ngamaba et al. [30] concluded that life satisfaction is preferable to happiness as a measure of subjective well-being. Future expectations were not related to happiness in this population (Table 4), contrasting with findings in the general Spanish population [31], although six participants in the qualitative study did relate happiness to both satisfaction with life and future expectations. In the present sample of HPs, twelve were classified with probable depression and ten with probable anxiety, but neither was significantly associated with happiness/unhappiness; however, there was a tendency towards an association with probable anxiety, which has been reported in the general population [32,33], and the failure to reach significance may be attributable to the small sample size. We cannot rule out the effect of other psychopathologies on the self-perceived happiness/unhappiness of HPs, and further research is required on this issue. The study by Cruz-Teran et al. [34] in Granada found a higher percentage of HPs who were satisfied with their lives. Besides differences in study design and measurement instruments, their survey took place before the economic crisis suffered by Spain and other European studies, which may contribute to this discrepancy. The authors concluded that the HPs in their study had little awareness of the reality in which they were immersed and denied their reality to strangers and those not living on the street as a survival strategy. No significant association was found between the responses to the question "Do you feel happy?" and any of the independent variables studied. Feeling happy was not related to age, sex, educational level, or housing situation, in agreement with the study by Panadero et al. [15] of 235 HPs. However, unlike this previous report, we found no association with a positive self-perception of health status in the first phase or future expectations in the quantitative study in the second phase, and health did not emerge as a happiness-conditioning theme in the qualitative study, although future expectations did, in various ways. These differences with the study by Panadero et al. [15] may be explained not only by the subjectivity of the participants' perception of their own health, but also by the different manner in which this information was gathered and the small sample size. Panadero et al. [15] reported a much more positive self-perception of health among those who felt happy rather than unhappy, and greater happiness was associated with a lesser degree of disability or severe/chronic disease, using a variety of questions and response categories. In contrast, the present participants responded to a single question on health with only two options (good or poor health). Panadero and coworkers [15] also described an association of happiness with religious belief in HPs, as in other populations [32,35], and HPs were found to score significantly higher for religious and spiritual beliefs in comparison to people with homes [7]. This aspect was not evaluated in our quantitative study and was not raised as a major theme in the interviews with HPs. Panadero et al. [15] found that happiness was not associated with having partners, family or friends, although the HPs felt happier "if they felt less abandoned". We did not gather data on this variable as such, but analogous expressions were frequent in the interviews, including: "being happy with the family", "seeing family and friends", "happiness of the group", and "happiness of others". The qualitative results obtained in the present study are in agreement with the conclusion of other studies [13,36] that happiness, life satisfaction, and subjective well-being are highly influenced by personal relationships (e.g., family, friends, acquaintances) and are associated with a concern for others, in line with previous findings in HPs [16,37]. Regarding self-esteem, more than half of the HPs in the qualitative study expressed feelings of satisfaction with themselves, similar to the finding by Cruz Teran et al. [33] that HPs generally have a rather optimistic perception of themselves; however, a minority of participants linked happiness to "having what is normal", "having a good lifestyle, a house, a job... ". The high percentage of HPs who reported feeling happy and satisfied offered various explanations for these states. Variables that were not considered in our study may also be relevant. For instance, a biological approach to happiness is gathering strength, and Panadero et al. [15] discussed the idea that optimism is a personality characteristic. Future research will shed light on the biological foundations of happiness and their implications, and a biological approach to positive emotions [38] and altruism [39] will also be taken into account. The HPs in our study did not tend to seek assistance from social services, despite generally being without unemployment, sleeping on the street, suffering assaults, and having a poor relationship with their families. These findings may help community nurses to better understand the alienation and mistrust felt by this stigmatized group and improve their approach to delivering care. Knowledge of what makes people in this situation feel happy and satisfied could be useful in the design of interventions to improve the health and well-being of the homeless. As noted by Vázquez-Souza [40], programs should be adapted to the places where HPs can be found and flexible to provide the continuous care that is required, including the treatment of disabilities, despair, fear, loneliness, and low self-esteem. Finally, there were much fewer women than men, as in other studies of this type [41], suggesting that women may preferentially seek informal solutions from friends and family members, contacting social services as a last resort. In addition, social services usually prioritize the provision of shelter to women sleeping on the street because of their frequent exposure to violence and sexual assault [42]. The large number of HPs feeling satisfied with their life and happy might suggest that many have a strong capacity for life adjustment, among other qualities, although this can only be considered as a speculative conclusion. --- Conclusions HPs who felt satisfied with life were 4.5-fold more likely to feel happy than those who did not. Happiness was related to material aspects, affective situations, and daily life activities, which they valued but did not possess. The majority of HPs felt good about themselves and half of them felt happy, despite their adverse situations, which included poor family relationships, no income, weak health, sleeping on the street, and vulnerability to assault and arrest. Knowledge of the feelings of HPs can help community nurses and other social care providers to approach this vulnerable population and can assist in the design of effective public health interventions to improve their physical and mental health. --- Study Limitations It is especially challenging to obtain honest responses to highly personal questions from this very fragile population. In order to minimize this potential limitation, the researchers spent two years establishing connections with HPs in the city on the street and in the Calor y Café center. Participants were selected by intentional sampling and only included HPs that were willing to participate and take part in a recorded interview. In addition, the results cannot be generalized due to the small sample size and the exploratory nature of the study, limiting the conclusions that can be drawn. A major obstacle proved to be the difficulty in contacting participants from the first phase of the study to take part in the second. Finally, further research is warranted to explore the impact of psychopathology on HPs' self-perceived health. --- Data Availability Statement: The underlying data are available on request. --- Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. --- Conflicts of Interest: The authors declare no conflict of interest.

Objective: To determine the relationship between the characteristics and experiences of homeless persons and their state of happiness as a basis for designing appropriate social support strategies. Design: Exploratory observational study with an analytical and descriptive qualitative design. Setting: Participants were contacted, administered with questionnaires, and interviewed in the street (central and northern areas of the city) or at the "Asociación Calor y Café" center in Granada (Spain) between April 2017 and February 2018. Participants: Selected by intentional sampling, 25 participants completed questionnaires in the first study and 14 of these were administered with questionnaires and interviewed in the second study. Method: General and specific questionnaires were administered to determine the state of happiness and other variables. Descriptive statistics were followed by an analysis of the relationships between variables and the content analysis of semi-structured interviews. Results: A feeling of happiness was described by 64% of participants and confirmed by a happiness scale score of 50%. Participants who felt satisfied with their life were 4.5-fold more likely to feel happy (p = 0.021). Expectations for the future were not associated with happiness or satisfaction with life. Content analysis of interviews revealed three main themes: conditions for happiness, own happiness/unhappiness, and self-esteem. Conclusions: Many homeless people describe themselves as feeling happy and satisfied with their life. Material aspects, affective situations, daily life concerns, and self-esteem predominate in their discourse on happiness.

. Introduction This mini review considers existing research on the emergence of the purposebuilt student accommodation (PBSA) market in the UK, focussing on England, and provides a benchmark from which further research can be explored to expand perspectives on densification, planning policies and impacts of student dynamics in cities. Since the beginning of the 21st century, students have become less reliant on traditional accommodation, historically dominated by university halls or houses of multiple occupation (HMOs) in the UK, as they have moved to embrace alternative, higher density living in private PBSA. Research on this sector, the implications of its emergence, and integration into UK university cities has so far been limited from planning perspectives when compared to academic literature from other countries such as Canada (Revington and Wray, 2022) and Australia (Holton and Mouat, 2021). PBSA as a built typology are often included in wider densification strategies and therefore contribute to increasing the local density in the neighborhoods in which they are developed. Churchman (1999, p. 390) defines urban density as a subjective measure that "can be perceived and evaluated in very different ways, by different people, under different circumstances, in different cultures and countries." How densification is adopted in policy and practice will depend on local contexts due to a variety of drivers rooted in social, political, environmental, and economic foundations. What is considered high-density development in York may not reflect high-density in London; it is quantifiable, but relative. In this mini review, we reflect on the way policies relate to density within the UK planning system, where "the often antagonistic, ambiguous nature of density and its application needs to be recognized by policy makers, regulators and involved parties... to contribute to the creation of successful places" (Livingstone et al., 2021, p. 203). In this instance, the creation of successful student living experiences in university cities. The PBSA market evolved rapidly in recent decades, reflecting large-scale private sector involvement in densifying student housing. PBSA today is a mainstream asset class for institutional real estate investors internationally (Newell and Marzuki, 2018), across the UK (Livingstone and Sanderson, 2022) and Europe (French et al., 2018). The UK is the second largest PBSA market globally, after the USA (Knight Frank, 2019). Of the UK's 2.2 million full-time student population, 1.63 m require accommodation (Cushman and Wakefield, 2022). The growth of PBSA emerged in response to continuing demand across the UK for alternative accommodation and increasing student numbers; it now provides 34% of all bedspaces (Carter Jonas, 2021). The PBSA market is highly regarded due to its resilience post Global Financial Crisis (GFC), and the fact it moves counter-cyclically to the economy works in the sector's favor, as student numbers continue to grow during this current period of post-covid uncertainty (Savills, 2021). Considering such growth, the PBSA sector has "had little influence on mainstream policy, particularly in the United Kingdom" (Revington and Wray, 2022, p. 3); this is the research lacuna we begin exploring. In this mini review PBSA is explored through a qualitative desktop methodology which examines existing literature on density and planning policies (Sections 2, 3). Due to the different devolved planning regimes across the four nations of the UK, our focus is on English cities and is grounded in existing policies. As noted by Revington and Wray (2022), there has been little work considering the connected forms of PBSA, density and planning in the UK market, and our conclusions provide provocations addressing this research gap (Section 4). ---. Density, planning, and PBSA in England Densification can be seen as an effective approach to addressing the UK's housing crisis, whilst supporting cities in becoming more sustainable (Breheny, 1996) and affordable (Aurand, 2010). Density is perceived by Holman et al. (2015, p. 2) as "a simple yet effective measure that brings together economic, environmental and social benefits solving the problems of a sprawling society, " however the authors recognize the nuanced nature of density, and it isn't a "winwin characterization" (Holman et al., 2015). Negative consequences of densification can be experienced through increased urban inequalities (Immergluck and Balan, 2018) and failures of property-led regeneration for local communities (Tallon, 2013). In exploring densification, de-densification and re-densification, McFarlane (2020) emphasizes that these processes are relational, resulting in urban transformations with varied characteristics. These relational densification processes are inherently local in context but are influenced by national regulation and governance, through planning policies. The UK's discretionary planning system has become increasingly deregulated, adopting a pragmatic and liberal social model (Nadin and Stead, 2008). Since the GFC, planning has seen additional reforms fundamentally driven by neoliberal logic (Lord and Tewdwr-Jones, 2014). The neoliberal turn led to financialization, concomitant to the increasing commercialization and internationalization of the UK university sector, and the emergence of PBSA as an asset class in the 21st century. Reynolds (2020, p. 2) suggests that "PBSA is a key example of rental market financialization" and we see a market dominated by institutional investors, such as real estate investment trusts (REITs, e.g., Unite) and global private equity funds (e.g., Blackstone). Fiorentino et al. (2020) began to unpack processes of financialization and density in the UK, suggesting that although the sector is experiencing large in-flows of international capital, caution should be exercised regarding market saturation. The National Planning Policy Framework (NPPF) provides the overarching policy guidance for England. It sets the directions to meet sustainable development goals and it is the foundation for plan formation across local authorities. Policy makers at local level have the capacity to interpret the NPPF using their own discretion relating to perceived community needs, such as housing targets. In leading university cities like London and Manchester the local housing target is driven by the demand and supply challenges of housing and affordability crises. When PBSAs are delivered within new developments the units count toward the local housing target, however, being temporary accommodation for students they do not effectively contribute to improving local housing provision nor affordability. In contrast, they contribute starkly to boosting local density figures. National Planning Policy Framework/NPPF (2021, p. 37) discusses density briefly, primarily concentrating on density for housing and new developments. It actively discourages "homes being built at low density, " suggesting minimum local density standards be significantly uplifted where appropriate. Density should be considered as flexible, when considering specific local areas, and "optimized" where possible. Therefore, the construction of PBSA units is frequently justified locally in connection with supply and demand dynamics in university cities, and within the remits of planning for subjective understandings of optimal density. The influence of students and universities are also undeniably connected to the increasing provision of accommodation through higher density PBSA. As student numbers continue to grow in the UK, the student experience is increasingly commodified (Chatterton, 2010). In this respect students are driving densification, supported by policy. The diversity of offering among PBSA providers has grown, however local city contexts and market dynamics remain indelibly important. In Liverpool, Mulhearn et al. (2018, p. 477) predict that the PBSA market will experience an "disorderly end" due to market saturation, speculation, and risks. Even with such risks, we suggest that PBSA development is further supported by viability assessments which often result in higher density outcomes and are exempt from National Planning Policy Framework/NPPF (2021, p. 18) affordability requirements (10% nationally), as they are providing "specialist accommodation for... people with specific needs". London is an example of how a locality has interpreted the NPPF to boost density for reaching housing targets and resulting in the largest market for the PBSA sector in England. As of 2021, London's 38 universities provided higher education for <unk>370,000 students, with 24% living in private student accommodation (Savills, 2021). In the current London Plan (GLA, 2021), the pre-existing "density matrix" which offered a defined approach to densifying, was scrapped in favor of optimal density with a requirement for affordable provision of PBSA. The expectation from the GLA that London will need 3,500 additional student beds p.a. (GLA, 2021) justifies the need for PBSA but recognizes the need for access to be more affordable. The plan defines affordable with a fuzzy sentence: a "PBSA bedroom that is provided at a rental cost for the academic year equal to or below 55 per cent of the maximum income... from the Government's maintenance loan for living costs" (GLA, 2021, p. 209). So, in the future, affordable provision should account for between 35 and 50% of the bedrooms. This prerequisite is likely to impact developers' viability studies, and further encourage "optimal" densification. ---. Too accommodating? Current PBSA policies Reflecting the complex dynamics active within the PBSA market, literature suggests that "policy responses are required" (Kenna and Murphy, 2021, p. 139) to address the variegated consequences of intensive and high-end PBSA development, as "there are limited regulations about the style, pricing, or other related aspects of these developments" (Kenna and Murphy, 2021, p. 151). Students may eventually choose to attend universities away from prime locations (e.g., London or Manchester), due to prohibitive costs, even with the affordability requirements of the London Plan (GLA, 2021). Literature confirms that the choice and availability of desirable accommodation is a key influence on students when deciding where to study (McCann et al., 2020). However, as discussed by Reynolds (2020), there is a student precariat who are unable to afford to live away from home, or higher rents typically associated with certain luxury PBSA. Providers may be reluctant to respond to a more equity-based policy driven approach, if viability and longer-term cashflows are potentially diminished, and shifts in the UK's discretionary system could be perceived as prohibitive. Mulhearn et al. (2018) advocate planning control for reasons more connected to market factors, oversupply, and the formulation of development models. University cities across England have begun to respond to the growth of PBSA in their localities through discretionary planning policies. In a number of locations, the impacts of PBSA growth are currently under discussion. In Sheffield, existing policies pertaining to universities and PBSA developments (Sheffield Core Strategy, 2009;Sheffield Student Accommodation Strategy, 2014), are viewed as outdated and disconnected with current educational trends. A report by Cushman and Wakefield (2021, p. 13) discusses concerns around PBSA oversupply, recommending that areas of the city are zoned for PBSA, that sites should only be released by the Council for PBSA in "exceptional circumstances, " that universities should play a more central role in such provision, with future planning policies more accurately incorporating supply and demand. Rooms in PBSA should offer more choice due to the preponderance of studios, with the report encouraging innovation in design (Cushman and Wakefield, 2021). The study notes that there is "little evidence that intensive levels of PBSA development are making inroads into the HMO market" with HMOs still very much in demand (Cushman and Wakefield, 2021, p. 13), therefore the proliferation of PBSA is not releasing housing to local markets. In Reading, the Reading Borough Council (2019) devotes a section to PBSA stating that new accommodation will be on or adjacent to the university, with a "presumption against proposals" (Reading Borough Council, 2019, p. 95) on other sites in attempts to address housing development needs. Unlike Sheffield, Reading sees evidence that student numbers living in HMOs have dropped when new PBSA units have been opened. Affordability concerns are not addressed, but the plan recognizes the need for flexibility depending on increases in student numbers. In line with London, York has been moving toward providing affordable PBSA. The city expects continuing growth in student numbers, and will encourage university agreements with providers, whilst limiting development on off-campus sites and encouraging brownfield regeneration (City of York Council, 2022). Interestingly their analysis accounts for one "low density" PBSA development, in a city where most schemes are of higher densities (Porter Planning Economics, 2022, p. 2). Concerns have been raised in relation to oversupply of PBSA beds (specifically studios, like Sheffield), sites being used for PBSA rather than housing, and the negative impacts caused by high concentrations of students. Future PBSA developments are likely to reflect on-campus densification where possible. Although many of the issues discussed in this section are locally contextual, they are also very similar. Local authorities are making concerted efforts to plan for PBSA developments, responding to the university needs within their cities. Although these are still in their infancy, they are reactive and ongoing, but perspectives on "optimal" density are not explicitly addressed. Could additional SPGs at local level, a supply and demand matrix for assessing student numbers, specific density and viability SPGs, or explicit affordability requirements through national policies help proactively address some of these challenges?. --- Conclusions: balancing intervention, densification, and interests The substantial development of PBSAs in English university cities reflects the impacts of increasing density through the interactions of the planning system, PBSA sector, and university sector. PBSA is a resilient investment that has generated consistent profit in the UK market post-GFC and is actively responding to locally specific, changing student dynamics, but also profiting from a discretionary planning system. This mini review has reflected how the discretionary nature of the planning system in the UK allows for an interpretative approach to regulations of PBSA in England, within the remit of "optimizing" density. We have seen how local authorities, such as Sheffield, York, Reading, and London have been responding to the growth of PBSA provision. Although their emergent approaches are considering how to best plan for ongoing developments, are the mechanisms being considered now a case of too little too late? The cumulative impact of higher density PBSA developments in particular parts of cities does not seem to be adequately considered by planning authorities. Given assumptions that higher density leads to more sustainable city form, this gap seems perplexing. Assumptions about activities that come from housing development (jobs, services, life on the street) appear to be na<unk>ve at best if the units are empty for part of the year and don't house residents. Moreover, PBSA do not present flexibility in being converted in other uses should the student demand fall. Both factors present severe risk to the future economic resilience of certain cities. A number of questions remain open. Even if housing targets are "achieved, " is PBSA density "optimal"? Is concentrating students in one particular area actually best planning practice due to the possible negative impacts of students in neighborhoods? How can affordability be better accounted for? How can we minimize negative impacts arising from PBSA market saturation and hyperdensification? What could additional research, on local city markets with dense PBSA developments reveal? Once again, because of the discretionary nature of the planning system, just like for housing targets, the overall density and number of units that developers manage to negotiate varies city by city. While writing this piece, the possible implementation of nationally defined housing targets is a contested matter in planning and political debates. If we take the case of UK PBSA, therefore an open question remains: who should have a say in defining unit targets and standards? Should universities have greater agency and influence in PBSA related planning and development decisions? More UK based evidence and research is needed to answer such questions, and approaches to PBSA internationally could be reviewed considering alternative responses to "optimal" densification. In conclusion, we suggest that planning authorities must better identify opportunities for PBSA based upon an understanding of neighborhood capacities and continual discussions with universities. By suggesting further research avenues, we again return to the idea of how we can better understand intervention, densification and PBSA impacts through balanced, timely, flexible and well-informed planning policies. As this mini review has illustrated, the somewhat disjointed and delayed responses to integrating PBSA density more definitively into national and local policies, has created the possibility for local market disruptions and disparities. --- Author contributions All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication. --- Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. --- Publisher's note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

This mini review explores perspectives on density and discretionary planning policy in relation to the emergent landscapes of purpose-built student accommodation (PBSA) across selected English university cities. It examines the nexus between density in planning policies and transformations in student accommodation, presenting a research gap requiring further investigation. Our research builds upon limited literature on UK PBSA within which there has been inadequate interrogation of policy impacts on the substantial growth of this now mature market sector. By better understanding the relations between national planning policies and their local interpretations regarding PBSA, we can move toward improving understandings for urban futures. The intersectionality between the evolution of the PBSA sector and the planning system are currently underexplored. The paper highlights the proliferation, densification, and concentration patterns of PBSAs and their connections to the surrounding planning system. The emergence of privately developed PBSA in the UK is typically characterized by medium to high rise, and medium to high density development. Higher density PBSA has created a dynamic spectrum of impacts reflecting subjective perceptions of their emergence in university cities. Our mini review illustrates the growth trajectory and direction of the PBSA sector. We conclude with a reflection on the discretionary nature of planning policies for density that often create regulatory loopholes allowing PBSA to thrive as a niche real estate asset class but not necessarily as a local and social collective asset.

Background There are processes and situations that define women's life paths and place them in extremely vulnerable scenarios. An example of this is homelessness. In recent decades, the study of the phenomenon of female homelessness has made it possible to ascertain the relevance of any form of violence against women in the lives of homeless women [1][2][3]. As adults, domestic violence may lead them to homelessness as they leave their homes to escape abuse [4,5]. In other cases, when women are unable to access housing, they are forced to endure abuse and violence [6]. In addition, women who remain on the streets are regular victims of rape, sexual assault, robbery, insults, and threats. Shelters for homeless people are not safe spaces for them, as they are then at risk of being physically or sexually assaulted by other residents [7]. Even though being a victim of certain forms of violence, such as intimate partner violence, makes it possible for women to access specific housing resources, sometimes women's biographies involve victimization accounts of certain types of "violence" for which there is a lack of a specific network of care. In other instances, their experiences of violence may remain invisible in the eyes of professionals [8], which is when their homelessness is seen --- Open Access *Correspondence: [email protected] 3 Nursing Department, Faculty of Medicine, Autonomous University of Madrid, Madrid, Spain Full list of author information is available at the end of the article as the primary problem and they are thus housed in specific residential facilities for this population. Homelessness has emerged as a growing concern in a number of countries, irrespective of the continent or the per capita income of cities, suggesting that the phenomenon may be a more structural and/or social problem. For instance, in Madrid, Spain, the most recent nightly survey of homeless people identified 2998 individuals living in shelters, shelters for immigrants or social housing, or sleeping rough. Of those 2998 individuals, 11.2% were women. According to the same survey, 55.1% of homeless people said that they had been victims of some type of violence on the street. Gender-based violence was mentioned by 2.3% of the women as one of the main reasons for sleeping on the street, by 3% in the case of women living in a shelter, and by 7.4% of women living in social housing [9]. A number of studies report the role of structural and interpersonal gender-based violence in the trajectories of homeless women [10][11][12] and the relevance of genderbased violence in their pathways to homelessness. Understanding and looking deeper into biographies shaped by violence and the effects of said violence on the different aspects of the individual is thus extremely relevant and necessary for the design of policies and strategies to be able to meet demands for care. However, it is important for research to be conducted in a sensitive manner to avoid revictimizing women and for it to promote their empowerment using a trauma-informed approach to research [13,14] and the emancipatory nursing praxis (ENP) as a nursing theory of social justice. In line with the considerations set out above, this case study aims to unravel the chaotic discourse of a suffering woman who is repeatedly subjected to gender-based violence. Furthermore, on a methodological level, this study aims to demonstrate the experience of working with photo-elicitation [15], a tool with great potential when it comes to guiding and facilitating discourse production in individuals in a situation of extreme vulnerability. This technique offers us the possibility to facilitate the production of meaningfully rich data while empowering participants and keeping them safe from revictimization. --- Methods --- Contextualization of the case study This case study is part of a larger study that aims to determine how homeless women perceive the violence they are exposed to and/or have experienced. To this end, a qualitative study was conducted under the theoretical proposal of symbolic interactionism and the methodological approach of grounded theory [16]. Maria (her name has been changed for anonymity) was recruited in the context of an observation program conducted using in-depth interviews in different shelters belonging to the municipal resource network of the Madrid City Council, Spain. Her account was identified as especially meaningful among the experiences of the 20 homeless women who were interviewed, exhibiting experiential characteristics relevant to the understanding of the phenomenon under study. It is for this reason that we set out to work from a single case study approach of an intrinsic nature. Case studies focus on the uniqueness and complexity of a single case, gaining an in-depth understanding of it. Qualitative case studies are thus characterized by their intense nature, providing a deep, dense, and holistic description of a specific phenomenon from a naturalistic perspective. In particular, we chose to conduct a unique case study of an instrumental nature, that is, the study of a critical or revealing case with its own suggestive characteristics that illustrates an exemplary situation. Our objective and reasons for doing so were, beyond simply understanding this individual's particular situation, to contribute to the understanding of a broader phenomenon and to refine theoretical proposals [17]. --- The participant One of twenty women interviewed, Maria provided highly valuable data to clarify and shed light on certain aspects, key for the overall understanding of the adaptive processes of homeless women who are victims of violence and their struggle to move on and become highly resilient survivors. Maria is a middle-aged woman immigrant with primary education who arrived in Spain fleeing a situation of domestic violence in which she had been suffering since childhood, leaving her two children in her country of birth. Since her arrival in Spain, she has gone through different scenarios and situations of gender-based violence that have led her to homelessness. Despite the consequences that the experiences of violence had on her physical and mental health, Maria was able to develop coping strategies to give meaning to her life and keep her hope alive. --- Data collection Data on Maria's narrative were collected in different stages. The first stage consisted of conducting an indepth interview in parallel with the interviews conducted with other participants in the study. An interview script was developed as a supporting element for the researcher (main author) (Table 1). The dynamic interview questions were designed based on the objectives, and the specific research questions were adapted to the participant's level of understanding. Special attention was also paid to aspects related to the amplitude versus the specificity of the questions and the sensitivity and intimate nature of their content. The interview had a duration of 100 min and was recorded on audiotape for subsequent analysis with Maria's permission. During data analysis, some meaningful elements emerged, inviting us to return to the field to delve deeper into and explore nuances, attributes and characteristics. One of these core elements that shaped Maria's experience was artistic creation as an expressive resource and as a tool to work on coping and resilience. We met with Maria again to invite her to participate in our proposal to collect data on artistic activities, which were important in providing her with motivation. Specifically, she was asked to continue to elaborate on her experiences through photo-elicitation. In this methodological proposal, first described by Collier in 1957 in the field of visual anthropology, photographic images are to be used as an evocative and provocative element of the informant's experience, stimulating deep emotional aspects of their experience and rescuing visual metaphors as expressive and discursive elements of great potential [15][16][17]. In addition, Wang and Burris [18], allude to this technique with the term "photovoice" in the context of community research. These authors understand that the potential of this technique lies in giving people a voice through photography in what constitutes a creative and empowering process for them. Through photovoice, the individual identifies, captures, and visually represents aspects of their environment and daily experiences using photography, which then makes it possible to analyze the problems, strengths, and ups and downs of daily life through critical and reflective dialogue. Previous studies have reported different experiences in the field where photography has been used as a provocative element in the sensitive discourse of individuals in particularly vulnerable situations, such as those of individuals with mental health problems [19], women who are victims of gender-based violence [20] and homeless women [21]. In this second meeting, Maria was told about the photo-elicitation technique, and she chose to use the camera on her own cell phone to take the pictures. It was explained to her that she could take pictures of things, objects, actions, places, or people (with their consent and without taking pictures or capturing elements that might reveal the person's identity) that she considered significant in relation to her experience of recovery, coping, and resilience. She was also asked to write a brief commentary on each of the photographs, briefly summarizing why she chose to take that photo and explaining its significance. Maria took the photographs over a period of four weeks, which consisted of a total of 24 pictures. She decided to print them herself and write her reflections on the back in her own handwriting. Once the task was completed, the principal investigator collected the iconographic material to prepare the subsequent discursive meeting. In order to prepare this second interview using the photographs as an evocative element, a script was made using one of the photographs as an example (Table 2). The script was designed by two researchers individually, who agreed on a single proposal based on 14 of the 24 pictures taken by the participant (Table 3). The interview was conducted in a meeting room of the shelter itself by the two researchers (main author and corresponding author) and had a duration of 100 min. During the interview, the interviewers sat on either side of Maria, keeping a distance of about 20 inches from her, and without any other elements in between (at a round table ). The researcher was to show the photographs in the order suggested by the participant's discourse, introducing the questions as appropriate. This task required a certain level of concentration and capacity of interrelation between the evocative resources and the evoked discourse. The second interviewer, on the other hand, focused on registering significant discursive aspects (as field notes) and introducing clarifying and explanatory questions in this respect. The interview was recorded on audiotape and transcribed in its entirety for subsequent analysis. --- Analysis The analysis of the discourse produced during the first in-depth interview was performed by the main author following the grounded theory proposal [22], that is, an open coding of the interview was performed by identifying significant elements, the meaning of which, in relation to the study objectives, was condensed by means of free codes and in-vivo codes. Each of these codes was defined to form a glossary of codes. Some of these comments on the codes gave rise to incipient interpretative memos when compared to the analysis of interviews with other women. Subsequently, all the codes identified were interrelated in the form of a network in order to build a guide map to further deepen our understanding of Maria's experience. This first phase of coding as assisted by Atlas.ti software (version 8.0). The analysis of the photographs and the discourse produced during the interview was performed in different phases, as shown in Table 4, by main and corresponding authors. Firstly, an iconographic analysis of the images was carried out from the point of view of the manifest content, that is, from an essentially descriptive perspective of the element photographed [16]. This description was then linked to the written reflections written by Maria on the back of the picture and the discourse she produced during the interview with the picture in question. Once the materials had been interrelated, we proceeded to work on the open coding, taking as a reference the previously prepared glossary of codes, incorporating new codes and redefining others that were pre-existing. Codes were regrouped to create semantic groups that, little by little, gave shape to those categories that summarizes Maria's process of coping, revealing the latent content of her discourse [16]. --- Ethical considerations This study was approved by the Clinical Research Ethics Committee of Autonomous University of Madrid, Spain, under file number CEI-84-1555. Maria's decision regarding whether or not to participate in the study was respected at all times and she was provided with all the information regarding the objectives of the research and the purpose, conditions, and possible benefits of her participation in it by means of an informed written consent form. Maria's privacy and the confidentiality of her data were safeguarded by assigning codes and fictitious names to the individuals mentioned in her discourse. The provided participant name is fictitious and other possible in identifiers such as age and country of origin have been avoided. In addition, iconographic data have been pixelated guaranteeing participant anonymity. Throughout the data collection process, special care was also taken to manage the emotional vulnerability of the participant through constant reflexivity and flexibility [23], by maintaining a sensitive attitude during the interview (using active listening, empathy) aimed at perceiving painful aspects and generating confidence and calm. In addition, the interview script questions, developed by themes and dynamic questions, were ordered using the "exploratory funnel" strategy, starting with questions of a more general nature and ending with the more specific and sensitive ones [24], supported by recovery models such as the "Tidal model", by Phil Barker and Poppy Buchanan-Barker [25] and the "Self-transcendence" model, by Pamela Reeds [26]. --- Quality criteria Different strategies were considered to ensure the internal validity of the findings in terms of reliability and verifiability. Thus, during the research process, we worked on triangulating the researchers, both for the planning of the interviews and for the analysis of the data. In addition, study reports have been periodically produced and have How does a woman survive on the street? How does a woman cope with adversity? been reviewed by an external expert. His recommendations, comments, and questions have served to delve deeper into some key aspects for their clarification. Once the data had been analyzed and the report on the findings had been prepared, it was presented to Maria, so that she could read it and judge how her experience was reflected in it. A sustained effort has been made to provide a detailed description of each of the processes undertaken in order to ensure the transparency of the research process and its methodological rigor. Finally, a critical and reflective attitude has been maintained at all times by the researchers as a result of their commitment to considering themselves to be research tools [23,27]. --- Results Maria's experience has allowed her to provide an example of and illustrate very concrete and significant aspects of the experience of gender-based violence, homelessness, and the process of coping with vulnerability as a homeless woman. Her discourse may be summarized in the categories explained and illustrated below: --- Living in a spiral of violence Gender-based violence in its different forms is present throughout Maria's biography. Maria was born and lived her childhood and adolescence in a dysfunctional family environment where family violence was a part of everyday life. In her daily life, the maternal figure is blurred, without any female figure representing a reference point. She refers to her mother as both a recipient and giver of violence, fulfilling roles of submission and domination at the same time, maintaining power relationships that, in themselves, structure, normalize, and perpetuate violent relationships. Maria grew up in a hostile environment where she was socialized in a culture of abuse and violence and internalizes some attributes and characteristics of the female gender role that are very much tainted and conditioned by these values and beliefs. Maria grew up in the shadow of a negligent upbringing model, with no emotional ties, attachment relationships, or solid network to support her. --- Then... among the members of my family everything was dirty because of drug trafficking, cocaine from drug trafficking; I was already being abused by my father, by my mother [...]. I was sexually abused in my childhood repeatedly by the lovers my mother brought home! She'd sneak them into the house behind my father's back... I remember that they used to touch my legs, my privates... up here as well (touching her breasts) [...].I was sexually abused many times in my childhood [...]. I do not have one nice memory, everything [I remember] is awful! During adolescence and early adulthood, as a consequence of this culture of violence against women, Maria was subjected to sexual violence, raped by people in her family environment, resulting in two unwanted Maria then decided to put distance between herself and the violence and flee from this scenario, leaving her two very young children behind. During her escape, she unknowingly entered a network of women traffickers where gender-based violence against women made her an "item of merchandise". For almost ten years, she was sexually exploited and, at the same time, she was forced to use narcotics to be able to keep up with carrying out sex work 24 h a day, due to being in an intensive and alienating production system. It was at this stage that sexual violence became linked to economic violence. Violence became the only element that linked her to her --- Photograph Back of the photograph: "When I was painting this picture, I was thinking of my daughter" --- Description(manifest content) The image is of a rectangular work of art, hanging on a light green wall. It is evident that it has been hand-painted with a paintbrush. There are two equally-sized flowers which occupy more than half of the work, one above the other, both colored with blends of blue and white, and with thin branches of the stems crossing each other, and with irregularly-shaped sky blue horizontally-oriented strokes. --- Analysis (Latent content) Code: "Unfulfilled motherhood". It refers to the frustration of not having played the mother's role from the participant's perspective (providing protection, building emotional ties). This failure to provide maternal care is interpreted by her as a vital failure that generates intense suffering Code: "Ripped-out children". It refers to the violent separation from children. This separation is a consequence of homelessness and is characterized by an abrupt and forcible separation. The term "ripped out" suggests the notion that children are part of the participant's identity as a mother. Separation is conceived not only as a physical separation, but also as a separation that leads to an acculturation of her children into their new hosting environment (cultural separation in relation to values, beliefs, ideas, customs, etc.) surroundings, trapping her in a dead-end spiral where violence feeds on violence, with the exercise of violence being the only way to survive this situation. [ Leaving the previous situation meant entering a situation of structural violence. In spite of having been able to report the situation she has experienced, Maria found herself without the necessary resources to support herself in her situation of real vulnerability, which is when her experience of homelessness and begging started. The streets became a place to live in and from which to live and, little by little, begging was replaced by theft, which eventually led to another scenario of violence: prison. Maria's discourse invites us once again to reflect on the structural violence (administrative-legislative) that relocated her to a new scenario where freedom was taken away from her. --- [...] Once on the streets, I started serving jail time, all because of petty theft, nothing violent [...]. It was all because of stealing for my sustenance, that's when I became a criminal. In prison, Maria found a hostile environment where symbolic violence was exercised by both the institution and her peers. According to her discourse, this last experience was the straw that broke the camel's back, overwhelmed her, and accelerated a series of health-illness processes in her physical and mental spheres. Far from being able to be considered, to a greater or lesser extent, a rehabilitative resource, prison ended up devastating Maria's identity, stripping her of all traces of dignity, self-recognition, and hope. Maria's entire life journey through violence manifests itself explicitly with signs and symptoms consistent with post-traumatic stress disorder. After serving her sentence, Maria took to the streets, where she once again faced homelessness, this time without intrapersonal or interpersonal resources and in a situation of extreme vulnerability and social exclusion. It is at this point that she entered the social support network through the street services and began her life experience in the shelter. Some time after being in the shelter, after intense work and important advances in the recovery process, Maria met her partner with whom she built a life project outside the shelter, as a result of her interaction with her peers in the shelter and her search to establishing emotional bonds. This relationship is understood by Maria as a bid to build a full and meaningful life, with the loved one as the core element. However, intimate partner violence emerged to complete the spiral of gender-based violence and sent her back to the streets, forcing her to return to the shelter, to the comfort and support of what she understands as her only family. [...] He approached me as a friend and I began to fall in love with him... I had never felt anything like that about anyone before... He asked me to leave [the shelter] and so we did... He had an income and we lived on it in a room... and we got married. [...].Over time, he began to mistreat me, to insult me... shocking jealousy and from there to the arguments, fights... hitting... and after two years, one day he threw me out of the room and I was back on the street with my luggage... --- Confronting vulnerability and violence Faced with a life full of violence and in a situation of extreme social vulnerability, Maria reflects on her strength and capacity to overcome obstacles. Different coping strategies are identified in her discourse such as: the redefinition of herself as a strong woman; the creation of new family networks; the strengthening of a maternal role in her own identity; and the nurturing of her spiritual wellbeing. The development of these strategies has allowed her to advance in her recovery process, that is, to advance in the redefinition of herself and in the recovery of her capacity of self-determination and ability to make decisions about her own life, so that she can achieve a sense of fulfillment, a sense of personal wellbeing and health, and of social involvement and good quality of life. Maria's discourse, however, does not allow us to identify procedural aspects that clarify when these strategies were developed or what factors may have facilitated or promoted their development. Instead, her narrative, which is viewed through a certain retrospective lens, touches on these resources in terms of results. Nevertheless, one can sense a decisive event that marks a change in her path, which is her leaving the streets and taking refuge in the shelter. --- [...] I think you have to have like a... how should I put it?! I can't explain it, be myself! To be, to continue with what I have to do, not being involved in prostitution, drugs... [Now I feel] that I am kind of free, I am free in spirit [...], that I can do what I want: eat, paint, sleep; nobody harasses me, I'm not on drugs, I have my freedom, I can go to church, paint, be in the shelter; I can talk, it's like... something that's very beautiful, it's something wonderful, something I don't know how to explain; I am now kind of free from where I was for so long; I was pimped for many years, forced to be with various men, I was like that back then and now I am free from anyone telling me to do this or that; when they tell me something here they tell me nicely, and, to me, that means freedom, I am finally free. --- Being a strong woman Maria understands herself as a strong woman, interpreting strength from the notion of vigor, vitality, drive, nerve, or energy. At the same time, in her discourse one can glimpse the idea of strength as a bastion or bulwark, that is, as a solid and strong space for defensive tasks. Without a doubt, the case of Maria metaphorically responds to the idea of a bastion within a defensive stronghold in the face of continued attacks on her person, the various forms of violence. --- [...] I say I don't know if I'm strong or if I've been touched by a star or... what has touched me? Tell me! [...]. I found it (see Table 3, picture 9) around Pr<unk>ncipe Pio and I brought it and kept it in my room [...]. Ah, yes, how beautiful, I don't know, this picture is impressive [...]. I see her as being very angelic, with a very angelic face, I see her as being natural [...], I consider her to be strong. In all that I do... here, for 24 years, I've kept holding on and I consider myself strong; I think I am strong; I don't know if I may be strong to other people, but I think I am strong. She recognizes a gift, a star, in this strength, and is aware of the other women who, having shared experiences similar to hers, have fallen by the wayside. She sees in feminine attributes and characteristics a great capacity for self-improvement and resilience with which she identifies and builds herself by rebuilding and strengthening her identity. --- New family networks In her discourse, Maria refers to the creation of new support networks and the establishment of new bonds, which she understands as her new family. Her process of re-affiliation through new social bonds became her main material and emotional resource for her subsistence and adaptation to daily life. The establishment of networks not only with other homeless individuals, but also with personnel from the shelter, people from neighboring areas, religious institutions, and support associations, allows her to improve her self-esteem and sense of belonging by reducing her isolation and feelings of loneliness. --- [...] My family couldn't be bothered to look for me or anything, I pulled through by myself and with the help of the shelter. Your family is where you live! This is my family! The shelter thus became her reference point, her emotional family, her home. Therefore, her home is interpreted as a relational space that provides her with protection against the dangers of the streets. Maria refers to the shelter as her formal adaptive psychosocial support network, where she has established special bonds with the center's staff and other users. Interaction with what surrounds her and those around her allows her to build new meanings around the reality she has come to live, by sharing beliefs and values, ways of doing and understanding things, which link her to the shelter, even to the point of considering it as "her home". The shelter, and her belonging to it, takes on meaning by becoming that which she never had at her side, her family. In addition, Maria also makes the space her own by decorating rooms with her works, which shows how our home shapes our identity and vice versa. Freedom is what I have always longed for, what I am looking for; I came here from my country looking for that freedom, to get out of the horror of my family..., I came here to bring order to my life, to see what a home really is, in spite of everything, and all this is the next best thing... paintings of mine that I have been doing as therapy to get out of my system everything that I had inside, all that anger, fear, pain, and I wear it like a tattoo so that I don't forget my country, to which I will never return, not even to meet my children. --- Re-building mother-child relationships For Maria, the fact that she missed out on raising her children generates frustration and pain. Her childhood experiences of lack of love and care became learned behaviors from her mother's role as a model of insecurity in her relationships and, consequently, deficiencies in her supporting role for her children. The loss of all contact with them over many years leads her to work on redefining mother-child relationships and maternal roles in order to justify a vital motivation and overcome "the pain of an unfulfilled role". She alludes to the fact that it is violence which has deprived her of the opportunity to become a mother using the expression "the children that have been ripped out". Maria feels motivated and hopes to establish emotional ties with her children and now grandchildren; to rebuild the emotional bond that was broken by time and space after years of separation. I --- Nurturing spiritual wellbeing Nurturing spiritual wellbeing by "seeking balance" refers to Maria's willingness to experience and integrate the meaning and purpose of life by connecting with the self, others, art, music, literature, nature, or a higher power than oneself, which can be strengthened. In her discourse, Maria alludes to the search for this wellbeing through various aesthetic and creative resources. Thus, she refers to having found in artistic creation processes a way to represent and externalize her pain and suffering. Creative processes thus became a key element in her recovery process, with great therapeutic potential. Writing, painting and photography are understood as projective elements that allow her to explore her emotions, perceptions, and behaviors from within her freedom, providing opportunities for self-knowledge and personal growth. When I paint I feel very happy, I forget about all my problems, my hand doesn't shake, I forget about terrible things [...]. I like to paint, to be calm, and to get out everything I have inside. This intrapersonal knowledge opens up one's consciousness to working on the spiritual dimension and encourages the development of new strategies and resources for self-transcendence. In Maria's case, moreover, religious beliefs and practices (reading the Bible, praying, volunteering in a church) coexist to enhance and promote this spiritual work, which gives her self-confidence and hope. (See Table 3, picture 1) I'd sit there and say: My God, I must have a balance! When I leave here I have to be stable and not get into drugs, alcohol, prostitution... God give me that strength! --- Discussion Maria's case study had a twofold purpose: to understand the experience of a homeless woman who was a victim of gender-based violence, which was considered to be a particularly representative case because of her ability to cope with adversity; and to generate a space for free research where the participant and the researchers could test new discursive codes to explore the experiences she had lived. Through photo-elicitation it has been possible to share a genuinely different experience, capable of expanding discursive possibilities when with words alone it is difficult to support the narrative of experiences because of their harshness and situations of extreme vulnerability [28,29]. Nevertheless, it is necessary to point out that this technique has required effort, time, capacity for introspection, as well as certain skills in the use of technical means from Maria. In addition, this technique has been expensive for the researchers themselves, since it has required a complex process of planning, monitoring, and analysis. Through this case study, Maria was able to describe numerous situations throughout her biography where violence (in all its forms and aspects) was present in various situations occurring consecutively, from childhood to mature adulthood. Homeless women are a very heterogeneous group and, as a result, so are the paths that lead them to homelessness as well as their later trajectories. However, there are circumstances, such as violence, that put them in a particularly vulnerable situation. Violence in the family environment during childhood is pointed out in numerous studies as a common factor for many of these women [30]. These experiences of violence have resulted in serious physical and mental health problems. Consistent with Maria's experience, the evidence indicates that post-traumatic stress disorder is the most frequently developed mental health disorder among victims of various forms of emotional, physical and sexual violence in both childhood and adulthood [3] and that this condition strongly influences the development of coping styles [31,32]. Cumulative exposure to stressful events, mostly related to violence (with special emphasis on domestic violence and childhood sexual abuse), is also a factor related to substance abuse. Homeless women refer to the fact that, sometimes, substance abuse is closely linked to the relational environment in which they live. On other occasions, there is a therapeutic justification for substance abuse, such as the self-management of anxiety and emotional distress [33]. In addition, we see how Maria has been in and out of homelessness at different times in her life. In line with this, Broll and Huey [34] conclude that homelessness tends to be cyclical, with women entering and leaving this situation. The results of the monitoring of a cohort of 269 homeless women suggest that this trend is greater among women who have experienced multiple forms of violence (e.g. physical or sexual abuse) during different stages in life (childhood and adulthood). In spite of the circumstances experienced and their impact on women's physical and mental health, women still seek and find motivation to continue to fight and move forward [30]. The latter can also be clearly seen in Maria's narrative. The identification and exploitation of internal resources, the strength she refers to, hint at coping from an adaptive approach, which is grounded in an exercise of restructuring and redefining one's own self and one's support environment [35]. The way in which she describes herself is from her encounters with others; it is from relationships that the individual is identified and reconstructed [36], tracing the path to move forward [37]. The lack of a family and affective ties makes Maria transform the shelter into her "home" by generating new networks that allow her social ties to be re-affiliated, thus reducing isolation and the feeling of loneliness. This experience is consistent with the findings of Groton and Radey [38], who mention the loss of primary social networks such as the family and violence in the relationships within one's social networks as key elements leading to homelessness. They also conclude that building new social networks is a powerful coping mechanism and that willingness to help peers generates a strong sense of wellbeing. Other studies also mention that living together in shelters enhances empowerment and self-esteem [39] and individuals interpret their experiences there as positive, particularly with respect to staff. In general, the shelter is a place of respite, providing safety and warmth after having been abused and socially isolated [40]. Domestic violence in childhood and intimate partner violence are significant psychosocial contextual factors that affect the regulation of emotions, fostering the development of post-traumatic stress and compromising future parenting [41]. In addition, learned helplessness [42] and the distorted maternal role model, with no reference figures or child attachment, shape the individual's future ability to provide safe parenting [43], defining a profile of vulnerability in one's role as a mother. Similarly, maternal identity is compromised by not fulfilling one's role as a mother [44]. The restoration of mother-child ties would not only allow her to learn about the fate of her relatives and children, but also to heal wounds and rebuild her relationship with them. These findings reinforce the importance of supporting maternal mental health and facilitating processes that allow women exposed to violence to recover [45]. Moreover, cultivating spirituality is another key aspect of recovery from trauma, focusing on oneself as a form of self-care, regaining control and empowerment [46] and God as a source of comfort and post-traumatic growth [47]. This echoes the findings of Ahuja et al., who conclude that levels of subjective wellbeing are lower among homeless people ("unhoused people") compared to housed participants, especially in dimensions such as social connectedness, negative emotions, perceived stressors, and resilience. On the other hand, spirituality and religiosity are more developed among unhoused people, who experience a more pronounced need to search for meaning and connect with a higher being [48]. Finally, the exploration and expression of feelings through artistic creation, in the case of Maria, show that art therapy based on creative processes helps to improve the therapeutic relationship, restore balance, and channel emotions [49]. This finding demonstrates the importance of using art as therapy to improve self-esteem and provide more positive future prospects [50]. Creative processes facilitate the construction of escape spaces where one is able to "switch off ", to "change reality", to "isolate oneself ", and they constitute a powerful strategy for the control of anxiety and emotional suffering. Through creative processes, emotions are identified, evoked, channeled, and transferred. At the same time, homeless people identify these spaces as safe and healing and as a tool for self-knowledge and for the transformation of their relationship with themselves. These are spaces that facilitate self-expression by recognizing oneself in one's creations, exploring one's own identity, and rediscovering facets of oneself that one did not think existed, but do exist [50]. The findings described in this study may have implications for both shelter policy and practice with respect to caring for homeless women, but also for addressing their mental health and social and family integration. To conclude,

Background: Violence against women places them in a vulnerable position with regard to homelessness. Although sometimes invisible, women's homelessness is a complex reality shrouded in dramatic biographies that should be sensitively addressed to avoid revictimization.With the aim of understanding the chaotic discourse of homeless women's experiences of violence, a qualitative single-case study was conducted using the photo-elicitation technique. Data were analyzed in accordance with grounded theory. Results: The participant's discourse could be summarized in the following categories: "Living in a spiral of violence", "Confronting vulnerability and violence", "Being a strong woman", "New family networks", "Re-building mother-child relationships", and "Nurturing spiritual wellbeing". Conclusions: Supporting homelessness women requires an approach that focuses on the prevention of re-victimization and the consequences of violence in terms of physical and mental health. Shelters are spaces of care for recovery and represent referential elements for the re-construction of self.

of feelings through artistic creation, in the case of Maria, show that art therapy based on creative processes helps to improve the therapeutic relationship, restore balance, and channel emotions [49]. This finding demonstrates the importance of using art as therapy to improve self-esteem and provide more positive future prospects [50]. Creative processes facilitate the construction of escape spaces where one is able to "switch off ", to "change reality", to "isolate oneself ", and they constitute a powerful strategy for the control of anxiety and emotional suffering. Through creative processes, emotions are identified, evoked, channeled, and transferred. At the same time, homeless people identify these spaces as safe and healing and as a tool for self-knowledge and for the transformation of their relationship with themselves. These are spaces that facilitate self-expression by recognizing oneself in one's creations, exploring one's own identity, and rediscovering facets of oneself that one did not think existed, but do exist [50]. The findings described in this study may have implications for both shelter policy and practice with respect to caring for homeless women, but also for addressing their mental health and social and family integration. To conclude, it should be noted that the reality of the victimization of homeless women requires approaches along the lines of "trauma-informed care", a philosophy of care that focuses on the recovery of women, taking into account their pathways of exposure to gender-based violence, emphasizing their physical and emotional safety, their recovery of self-control, their strengths versus their weaknesses or deficits, and, ultimately, strengthening their capacity for resilience and empowerment [51]. --- Conclusions Gender-based violence, in its different aspects and representations, is a constant in the trajectories of homeless women. Welcoming and supporting them requires an approach that focuses on the prevention of re-victimization, as well as on the consequences of this violence in terms of physical and mental health. Shelters emerge as spaces of care for recovery and represent referential elements for the re-construction of bonding relationships, the creation of supportive environments, and the meeting point of development spaces in freedom and the capacity for self-determination. Sensitive, trauma-informed research should be used to inform the support offered to homeless women as part of narrative advocacy-based care. --- Availability of data and material The complete (de-identified) dataset generated and analysed during the current study can be made available from the corresponding author on reasonable request. --- Abbreviations MSc: Magister Scientiae (Master of Science. Master Degree); PhD: Philosophiae doctor (Doctor. The highest University Degree); PhD Candidate: Philosophiae doctor candidate; Psyc: Psychologist; RN : Registered nurse. --- Authors' contributions --- Declarations Ethics approval and consent to participate The Clinical Research Ethics Committee of Autonomous University of Madrid, Spain, under file number CEI-84-1555, approved this study. The participant was informed and voluntary gave written consent to participate in the study. --- Consent for publication Not applicable. The provided participant name is fictitious and other possible in identifiers such as age and country of origin have been avoided. Images have been pixelated guaranteeing participant anonymity. --- Competing interests The authors declare that there is no conflict of interest. • fast, convenient online submission • thorough peer review by experienced researchers in your field --- • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year --- • At BMC, research is always in progress. --- Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research? Choose BMC and benefit from:? Choose BMC and benefit from: --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: Violence against women places them in a vulnerable position with regard to homelessness. Although sometimes invisible, women's homelessness is a complex reality shrouded in dramatic biographies that should be sensitively addressed to avoid revictimization.With the aim of understanding the chaotic discourse of homeless women's experiences of violence, a qualitative single-case study was conducted using the photo-elicitation technique. Data were analyzed in accordance with grounded theory. Results: The participant's discourse could be summarized in the following categories: "Living in a spiral of violence", "Confronting vulnerability and violence", "Being a strong woman", "New family networks", "Re-building mother-child relationships", and "Nurturing spiritual wellbeing". Conclusions: Supporting homelessness women requires an approach that focuses on the prevention of re-victimization and the consequences of violence in terms of physical and mental health. Shelters are spaces of care for recovery and represent referential elements for the re-construction of self.

Introduction Researchers have long observed that some level of ethnic segmentation occurs amongst low-skill roles in labour markets where large numbers of migrants are received (see, e.g. Massey et al., 1993;Sassen, 1996;Wills et al., 2010). However, whilst this is occasionally due to a specific and sustained attempt by the receiving country to attract and limit immigrant workers to certain roles (Stalker, 2000), it is more commonly a phenomenon that appears to have no bureaucratic roots. In the language of complexity studies, it is emergent -that is, it arises from a concatenation of loosely related events and processes which, moreover, could not have been predicted from initial conditions alone. In this article we present findings based on an agent-based model (ABM) developed from the network-based, immigration-driven description of ethnic labour-market segmentation given in Waldinger & Lichter's (2003) highly cited study, How the Other Half Works. Whilst the notion that social networks influence the development of ethnic segmentation in low-skilled jobs has been widely embraced amongst migration scholars (see, e.g. Boyd & Nowak, 2012), it has proved difficult to analyse how this occurs in varying locations, and the strength of this causal process in comparison to other factors shown to influence the development of ethnic segmentation, such as discrimination by employers. Employing a sociologically informed ABM allows us the opportunity to develop an in vitro model (Edmonds, 2013) to test the mechanisms and relative strength of social networks as a casual factor behind segmentation in low-skilled labour markets. <unk> 4 <unk> Whilst many academics have forwarded empirically-based theories as to why ethnic labour market segmentation may arise, the interconnected and overlapping nature of the non-linear causal processes involved makes it extremely difficult to begin to uncover the extent to which such proposed theories can be usefully extended to help us understand similar situations in different times, spaces and places. Approaches that have attempted to explain labour market segmentation by trying to'reverse engineer' the processes involved, such as those using regression statistical modelling techniques, are restricted in their explanatory scope due to their linear nature. Whilst such techniques can be extremely useful to the migration scholar, particularly when explaining the strength of relationships between independent variables (such as ethnicity, nationality, age, level of education or gender) and a dependent outcome (for example, wage level, or the'segmentedness' of sectors of employment), they are only able to theorise as to why such a relationship may have developed and how local conditions may have influenced the development of such an association. In essence, they can provide a broad retrospective snapshot of the relationships between variables, but struggle to explain any emergent causation (see, e.g. Sorensen, 1998). Equally, in-depth qualitative and ethnographic studies of labour market segmentation have provided illuminating examples of how and why such processes occurred under the specific conditions under investigation, yet the very depth they provide limits the potential breadth of any explanatory power they may produce. In essence, they provide a very detailed picture of the functioning of the processes involved in labour market Immigration, social networks and the emergence of segmentation <unk> 5 <unk> segmentation within a very constrained time-space, yet struggle to generalise those findings to similar processes in other times and spaces. The use of agent-based models (ABMs), however, represents a shift not only from solely investigating variables and factors that might impinge upon a situation, but also from exclusively concentrating on a single in-depth case study. Instead, a consideration of socially-embedded individuals is produced, which allows for an experimental analysis of the interaction between the processes that act upon those agents. In this way, it combines elements of both quantitative and qualitative approaches to social phenomena. Furthermore, ABMs are particularly well suited to the study of complex social phenomena (Gilbert & Troitzsch, 2005). Complex systems are neither inherently orderly nor intrinsically disorderly but switch between both without solely being explained by either. Indeed, many social systems, including international migration (Papastergiadis, 2000) appear to fit into such a typology in that they exhibit key qualitative and phenomenological elements of a complex system (see, e.g. Geyer & Rihani, 2010;Mol & Law, 2002;Sawyer, 2005). Agent-based modelling is a simulation approach in which discrete computer generated 'agents' are imbued with limited autonomy, a capability to adapt or react to the environment, and are involved in local interactions. In this way, 'agents' in ABMs have agency in a way that is very familiar to migration scholars. Whilst this artificial agency is bounded by the simple rules of engagement programmed into the model, the interplay of discrete agents following their own sets of rules can produce non-linear outcomes on <unk> 6 <unk> a 'population' level. In this way, ABMs allow us to explore the same processes as existing sociological methods in a novel way. Furthermore, this novelty allows us to investigate non-deterministic social systems 'from the bottom up' in a way which is simply not possible using existing methods. For Edmund Chattoe-Brown ( 2013) such an approach is highly useful for social science because it presents another 'lens' on society, reframing existing sociological questions and allowing for a very different conceptualisation of causation. Whilst such an approach should not be considered a replacement for existing methods, it does provide both a useful addition in itself and a way to combine existing data (quantitative or qualitative) in a novel manner. the labour market was less volatile, but less inequality and smaller social networks in more volatile markets. However, we must remember that, in this model, maintaining costly social ties is an entirely rational economic choice. Calvo-Armengol & Jackson ( 2004), meanwhile, have illustrated how social networks positively correlate with employment patterns across time and between groups, but that when information about job vacancies spread through such networks, unemployed agents find themselves increasingly 'locked out' of employment as their term out of work increases. Furthermore, in a later article, the same authors (2007) observed that initial differences in wage levels and employment status are maintained over time for groups with identical social network structures. Krauth (2004) shows that, under conditions of referral hiring, an agent can increase their chance of getting hired by enlarging their social networks, particularly with ties to those already in work. Additionally, the work of Tassier & Menczer (2008) has revealed that, whilst random social networks increase one's likelihood of hearing about a vacancy, that information is also likely to be widely known. Where social networks are non-random, job information is shared with a smaller number of agents and employment rates for that social network increase. Abdou & Gilbert (2009) have modelled the relationship between social and workplace segmentation with reference to Egypt, illustrating how high levels of referral hiring lead to an increase in both forms of segmentation, particularly where there is a smaller minority population, or in smaller firms. Crucially, however, they recognise that, in <unk> 8 <unk> terms of overall employment rates, 'increasing levels of referral hiring could be beneficial for minority groups when the population is highly segregated and harmful otherwise'(2009: 19). Whilst these models deal with simplified scenarios compared to the rich data sources often utilised by sociologists, they do illustrate some important emergent properties of networked labour markets. Firstly, the structure of social networks matters for both an individual agent's chance of getting a job, but also for the development of populationlevel phenomena, such as segmentation. Secondly, the relative size of a minority population, the size of firms involved and their propensity to favour referral hiring techniques all contribute to the extent and type of segmentation, which occur. Indeed, for those familiar with Waldinger and Lichter's work on ethnically segmented labour markets in 1990's Los Angeles, these phenomena are recognisable; they stress the importance of homophilic social networks to getting a job, but also note that differences in employers' preferences for referral hiring affects the make-up of a workforce (2003: 97). This kind of ethnic segmentation, complete with referral hiring and self-regulation, has been widely reported in more recent studies investigating'migrant labour markets' (Bauder, 2006;McDowell et al., 2009;Wills et al., 2010) 1) Networks provide information -'telling job seekers about opportunities and informing employers about the characteristics of applicants'. 2) Networks are instruments of influence -'allowing job seekers to put themselves on the inside track by proxy'. 3) Networks can be used to enforce obligations -'so that the employer is assured that the favors he or she does for the job-seeker and his or her accomplices will be repaid'. 4) Networks can cement implicit contracts -because networks are carriers of information and obligation, they can be used to impose the 'rights and responsibilities of each party of the employment exchange.' Furthermore, '[t]o the extent that a group of workers feels bound by these understandings, the employer can count on its exercise of social control to keep recalcitrant fellows in line' (ibid). In the model described in this paper, we only model the flow of information and influence through networks. Obligations and implicit contracts, though important to a sociological understanding of referral-based network recruitment, are based on implicit systems, which were not evidenced enough in the study for us to convert into explicitly modelled functions. It can be argued, however, that giving employers a preference for referral hiring over hiring applicants who just 'walked in' off the streets is an implicit representation of the obligation aspect. Immigration, social networks and the emergence of segmentation <unk> 11 <unk> The above conceptualisation suggests two types of agent to be modelled: 1) Workers -they look for jobs and are enmeshed in social networks. They can pass information about vacant jobs through these networks, and also get a boost to their likelihood of getting a job they have heard about through such a connection. 2) Employers -they give out jobs and use the social networks of their existing workforce for recruitment. They have a preference for recruiting workers in the same social networks as their existing workforce. Additionally, we need to represent: Social networks -these connect groups of workers. They provide information about available jobs at the employers of social network group members. They also provide an advantage to getting a job where other network members are already employed in an organisation (influence). Another reason for agents to leave the modelled labour market is 'discouragement' (see, e.g. Herzog, Schlottmann & Boehm, 1993). This will occur if they fail to get a job for a set period of time (1 year in the experiments described in this paper). It is modelled as a probability of leaving (set to 10%), which is tested every time-step once the crucial time Immigration, social networks and the emergence of segmentation <unk> 13 <unk> period for an agent has been reached. Both 'native' and 'immigrant' agents can become discouraged. Our aim in designing the model has been to incorporate as much available data from the case study as possible, while still keeping the model as simple as possible. Due to the latter design concept, the labour market is restricted to low-skill jobs with no entry requirements; therefore skills of workers are not represented in this model version. To an employer, applicants for a vacant job are indistinguishable but for their ethnicity and number of links they have to the existing workforce. Similarly, workers do not have preferences for a particular job at a particular firm; to them all jobs are the same and they will apply to any firm one of their employed friends will tell them about. The data we could derive from Waldinger and Lichter's study is used to initialise both the labour market and the initial population of worker agents (see section 4.4). It has also been applied in calibrating the model so that its overall population dynamics correlate with the trajectories of the numbers of unskilled workers per ethnic group as reported in the study. This was necessary as no data regarding actual, historical immigration rates (i.e. including illegal immigration to the case study area) was available. --- Model entities The model is implemented in NetLogo (Wilensky, 1999). It distinguishes two types of agents, realised as breeds in NetLogo: 1. Workers: Worker agents represent individual people seeking employment in the low-skill labour market. They are characterized by their age (18-65), sex <unk> 14 <unk> (male/female) and ethnicity (an integer between 0 and number of ethnicities -1, which in this case is assigned as 0-White, 1-Black, 2-Latino and 3-Asian). Each worker agent also keeps tabs on the job they have (attribute job, set to -1 if unemployed), how long they have been unemployed for (ticksunemployed) and a list of current vacancies (job-offers) they know of via their friends. --- 2. Employers: Each employer agent represents a firm in the model. Firms are characterized by their type (one of six, attribute industry-index), their size (number of low-skill jobs, num-jobs), and their yearly turnover rate (annualchurn-rate). These are all taken from the available case study data (see section 4.4). Employer agents keep tabs on which of their jobs are vacant or taken (vacant-jobs / filled-jobs), who is currently applying for a job (applicants / walk-ins -the former have heard about a vacancy via their friendship network, the latter have not), and recommendations of their existing employees for particular applicants (referrals). For ease of producing model output measures, employers also keep track of how many workers of each ethnicity they currently employ (ethnicity-counts). In addition to these type-specific attributes, each agent has a unique id (attribute who, automati-cally assigned in NetLogo) and a location (given as the x-y coordinates of the grid cell they in-habit: xcor, ycor). At the beginning of each tick, internal variables are updated. This includes the age of workers (if the tick marks the beginning of a new year), which in turn may cause some workers to retire and leave the model. The main model processes are then carried out in the following order (with sets of agents always activated in random order): 1. Advertise vacancies: Employers 'advertise' any vacancies by notifying their employees. --- 2. Spread the word: Employed workers tell their unemployed friends about any vacancies they know of. They also recommend them to their employer, i.e. send a "referral" for each unemployed friend. --- 3. Apply for job: Unemployed workers apply for jobs. If they know of any vacancies, this means they send applications to the employers in question. Otherwise, they walk in at the nearest employer with vacancies to ask for a job. The model assumes here that a time step of a week is enough for any unemployed worker without ties to the employed workforce to complete such a search successfully (if there are any vacancies). --- 4. Screen applications: Employers pick a new employee for each of their vacancies (there can be more than one at a time) according to the specified recruitment strategy (model parameter e-selection-method). In the standard Immigration, social networks and the emergence of segmentation <unk> 17 <unk> model described here, this will be 'normal' referral hiring, i.e. the applicant with the most recommendations from existing employees will be chosen, regardless of their ethnicity 1. With 'internal' bias, employers strive to keep the ethnic composition of their workforce as it is, whereas with 'external' bias, they strive to match their workforce to the ethnic composition of the population (see section 5.2). --- 5. Leave discouraged: Unemployed workers update their unemployment time (attribute ticks-unemployed). If they have reached the threshold of time they are willing to wait until they get a job (model parameter ticks-discouraged), they may leave the model with a specified probability (model parameter prob-discouraged). --- 6. Job churn: Employers check if some staff turnover is happening in this tick, determined by their industry's annual turnover rate (attribute annual-churnrate). At the end of each tick, demographic processes occur. These are immigration for the non-native ethnicities (Latinos, Asians) and 'births' of 18-year-olds for all ethnicities. In addition, worker agents may grow their social network. All model processes are described in more detail in the model documentation available online with the model code (Meyer, 2018). --- Initialisation The model is initialised using the available data from the case study (Waldinger & Lichter, 2003: 235-252). The characteristics of the firms forming the labour market are <unk> 18 <unk> determined by distributions derived from tables A2-A9, with the values shown in Table 1. [Insert Table 1 about here] From this, the number of firms, and their initial vacancy proportion (model parameters num-employers and initial-vacancy-proportion) the modelled labour market is constructed. Locations of firms are assigned randomly. To be able to compare different simulation runs for the same labour market, a particular setup may be read in from a text file. The model starts from a state where only native workers are employed whereas all immigrants already present are unemployed. The total number of agents and their ethnic composition can be calculated from the number of initially filled and vacant jobs, the initial unemployment rate and the ethnic proportions (given by the model parameters initial-unemployment-rate and eth-proportions). For the experiments reported in this paper, the latter have been derived by linear regression from the case study data (Waldinger & Lichter 2003, Tables A2, A4-A8), which specify the working population in % for the different industry sectors and ethnicities for the years 1970, 1980 and 1990. The model initialisation uses the values for 1960 shown in Table 2. [Insert Table 2 about here] The model world is resized to easily fit the number of agents before creating all employers and workers. The employed proportion of native workers is then hired by the employers making sure that Whites and Blacks are evenly spread over all firms, i.e. that each firm ends up with a similar ratio of Whites to Blacks. Immigration, social networks and the emergence of segmentation <unk> 19 <unk> Finally, the ethnically separated social networks are initialised. Each worker forms some links with other workers of the same ethnicity in their vicinity. The number of links is determined by taking a sample from a random exponential distribution with mean given by the model parameter mean-initial-links, and then taking the minimum of this sample and the number of co-ethnics in the Moore neighbourhood of depth 5 of the respective agent. Links are undirected, therefore having one agent form a link with another agent means that both of them gain a 'friend'. Agents may thus gain additional links to their own initially formed links, resulting in a highly skewed degree distribution as is typical for social networks. --- Calibration We calibrated the model with a small labour market setup of 11 employers, offering a total of 2103 jobs, and the ethnicity proportions given above. Birth rates per ethnicity were taken from official statistics for 1990 (Livingston & Cohn, 2012), using the values for native-born Blacks, foreign-born Hispanics and foreign-born Asians (8.62%, 13.69%, and 7.85%, respectively). The value for native-born Whites had to be adapted so that the overall population dynamics of the model correlates with the trajectories of the numbers of unskilled workers per ethnic group as reported in the case study data (see section 4.4 above). In the absence of empirical data on actual, historical immigration rates into the case study area, immigration of the non-native ethnic groups ('Latino' and 'Asian') in the model is driven by the success of co-ethnics already in the labour market. The lower the unemployment rate, the higher the immigration rate. However, once employment falls <unk> 20 <unk> below a certain level (90% during calibration), immigration will cease until that level is again exceeded. This is subject to a 2-tick information delay (see, e.g. Stalker, 2000). Calibration resulted in the birth rate for Whites set to 0.5% (this acknowledges the fact that the White model population is mostly moving out of the low-skilled labour market over the modelled period of 30 years). We also derived fixed immigration rates for both non-native ethnicities for each year, with the initial vacancy proportion set to 10%. These rates have been applied in subsequent simulation runs with a larger labour market of 100 employers and a total of 5000 jobs, to make comparison between different model settings easier and take the influence of varying immigration rates on model results out of the equation. --- Results and findings Our concern in this paper was to model Waldinger & Lichter's How the Other Half Works with particular reference to the impact of homophilic social networks on processes of labour market segmentation in low-skill jobs where formal entry requirements are negligible (Bauder, 2006;Piore, 1979;Sassen, 1996). Therefore, we focus on i) the conditions in which ethnic segmentation occurred in the model, and ii) the relative impact of homophilic social networks under conditions of two forms of 'conservative' discrimination by employers; that is, employers seeking to match their workforce with an imagined customer-base -in this case, the total population of the model (Waldinger & Lichter, 2003: 131) -or to maintain the ethnic 'balance' of the firm, in order not to disrupt the delicate bonds of influence, obligations and implicit contracts within the organisation (ibid: 83). Immigration, social networks and the emergence of segmentation <unk> 21 <unk> All model scenarios use the same configuration of firms (100 employers with a total of 5000 jobs); all experiments were run for 30 years (1560 ticks). Results depicted in this section are usually averaged over ten simulation runs per scenario to account for probabilistic variance, if not stated otherwise. --- Segmentation measures To measure relative segmentation we apply a multigroup segregation indicator that is able to take several ethnic groups into account instead of measuring segregation between just two population groups, as is standard in segregation measures. Reardon & Firebaugh (2002) H =!! p!!!!!!!!!!!! r!" ln r!" (Eq. 1) with Theil's Entropy index E = p! ln (1 p!!!!! ) applied as the disproportionality function. The ratio r!" = p!" p! expresses the extent to which ethnicity m is disproportionately represented in firm j : r!" <unk> 1 indicates m is underrepresented; r!" > 1 indicates m is overrepresented in firm j. <unk> 22 <unk> The H index can be interpreted as 'one minus the ratio of the average within-unit population diversity to the diversity of the total population' (Reardon & Firebaugh 2002:42). The closer the average diversity within firms matches the diversity of the total population, the nearer this ratio gets to 1, therefore the closer the H index gets to 0 (no segmentation). The more the diversity within firms diverges from the diversity of the total population, the higher the H index (higher segmentation). This allows us to easily compare two different simulated scenarios and establish which one leads to relatively higher segmentation. In addition, we apply indicators we derived from the case study itself. To be used as segmentation indicators, we define a lock-out to occur when the proportion of native workers in a firm is greater than 80%; similarly, a take-over has happened when the proportion of immigrant workers in a firm is greater than 80%. To avoid overshadowing of the indicators by the population dynamics in this low-skilled labour market, where a White majority is slowly displaced by Latino immigrants, whereas the proportion of Blacks remains steady at around 10% of the population (see data in section 4.4), we decided to measure the number of lock-outs and take-overs not at the end but after 20 years of simulation (tick 1040). --- Impact of social networks on segmentation In Waldinger & Lichter's theorisation, social networks are of central importance to the development of ethnically segmented labour markets -they allow for information about job vacancies in niche areas to flow rapidly within ethnically homogenous networks where members are already employed within that niche, meaning certain groups could react more quickly than others to opportunities. Employers, meanwhile, could utilise these networks as a cost-effective way to gain employees with pre-existing obligations, not only to their employer, but also their co-workers and fellow network members. However, the degree to which such networks impact on the extent and rate of segmentation is difficult to specify without a control where such networks do not exist. Whilst this is not possible in'real world' research, one use of agent-based models is running just such a 'null' hypothesis and compare it to a'standard' model. In the standard model described in section 4, employment processes are influenced by both homophilic social networks (workers receive information about vacancies, employers receive recommendations for applicants from existing employees) and <unk> 24 <unk> spatial effects (unemployed workers without useful ties apply at the nearest firm with vacancies, new agents try to'settle down' next to other agents of their own ethnic group). To test the impact of these, we compared this model with model variants without one or both of these influences. In the so-called 'null' model, agents do not have social networks and space is basically reduced to one location so that all distances between firms and workers are identical. The third model variant lies in between: Agents have social networks but space is not modelled. As can be seen from Figure 2, the impact of the social network on labour market segmentation is substantial. While there is no segmentation in the null model, both the standard model and its variant without space show significant segregation levels. The H index for the standard model stabilises at a value of 0.4, about 30% higher than that for the model without space but social networks. [Insert Figure 2 here] This implies that, whilst they are not the sole cause of segmentation in the model (the geographic proximity of agents to particular firms and each other also plays a part), ethnically homogenous social networks substantially increase the overall level of ethnic segmentation within the model. Furthermore, in certain scenarios, such segmentation can become extreme, particularly for groups with a small (initial) cohort. Whilst system variation is to be expected, it is clear that introducing ethnically homogenous social networks increases the propensity for smaller groups to become highly concentrated within certain sectors of the labour market. In the case shown in Figure 3, this is true for both natives and immigrants as the two smaller ethnic groups (native 'Blacks' and Immigration, social networks and the emergence of segmentation <unk> 25 <unk> immigrant 'Asians') are excluded from more than 75% of firms at the end of the simulation with the standard model. [Insert Figure 3 of labour, social and cultural distinctions mark certain workers as more suitable for certain kinds of jobs (Bauder, 2006;Bourdieu, 2007;Wills et al., 2010). However, what form this discrimination takes has the potential to radically change the way ethnic segmentation in the labour market occurs. In this paper we test two forms of 'conservative' discrimination; in the first part of the experiment we assume that employers wish to conserve the current make-up of their workplace in order to preserve the functional order of the organisation. Only where the maintenance of the current ethnic balance of the workforce is impossible will the employer look further afield. We term this internally conservative discrimination (ICD). In modelling terms, the ethnic proportions are translated into probabilities of selecting a particular ethnicity to filter the pool of potential candidates. This means that if an organisation employs 75% 'White', 10% 'Black and 15% 'Latino', and there are one 'White' applicant and two 'Black' applicants for a job, there is a 75% chance that the 'White' applicant is chosen. If there are several candidates of the pre-selected ethnicity, the one with the most recommendations will gain the job as before. In case there are none, recruitment follows the'standard' referral hiring process. Thus any candidates <unk> 26 <unk> from ethnic groups not yet present in a firm will only get a role in this organisation if no other candidates apply. In the second part of the experiment, we assume that employers seek to build a workforce that reflects their 'customer base' (Waldinger & Lichter, 2003: 199). We describe this as externally conservative discrimination (ECD). Whilst this may be a more relevant form of discrimination in roles that involve a direct engagement with the public, such as restaurants and service jobs in hotels and hospitals, it provides a subtly different form of discrimination. Here, we modelled this process as employers having a bias towards hiring workers who would help their workforce reflect the overall population diversity of the model. Therefore, if we take the scenario described previously, where an organisation is composed of 75% 'White', 10% 'Black' and 15% 'Latino', and add the caveat that the overall population is 60% 'White', 10% 'Black', 25% 'Latino' and 5% 'Asian', then the probability of each of the four ethnic groups being selected to choose a candidate from, notwithstanding social network recommendations, is 0.6, 0.1, 0.25 and 0.05, respectively. If there is no candidate of the pre-selected ethnicity, normal referral hiring resumes. We implemented both forms of discrimination in the'standard' model and also the 'null' model, in order to be able to gauge the effect of homophilic social networks in these scenarios. Figure 4 shows results for both model variants and the two forms of discrimination. It also includes the previous scenario, the standard referral hiring (leftmost bars in Fig. 4), for comparison. hard for immigrants to 'get a foot in the door' when employers try to maintain the existing ethnic composition of their workforce. Only the existence of an initial surplus in vacancies makes it possible for immigrants to gain employment. But once a successful toehold is formed, their chances increase and social network processes will tend to reinforce that nascent advantage, as is evident in the higher number of take-overs in the standard model compared to the null variant. With the second form of discrimination, when no social networks exist, there is negligible ethnic segmentation (Figure 4, right). This is unsurprising as the ECD scenario means that employers actively seek to match their workforce to the overall population. However, despite this tendency, ethnic segmentation does occur when social networks are present, though this is far less pronounced than in the ICD scenarios, <unk> 28 <unk> and entails a reduction from non-discriminatory scenarios (the average number of lockouts is about halved, whereas take-overs are nearly completely eliminated compared to normal referral hiring, see Figure 4, left). Again, therefore, social networks have the impact of buffering migrant groups against the negative effects of discrimination, whilst increasing the propensity for such groups (especially if they are small) to become 'ghettoised' in the labour market. However, of perhaps greater significance, are the notably different impacts of internally-and externally-conservative discrimination regimes. Whilst both initially have negative impacts for immigrant incorporation into the labour market, once migrants begin to gain a foothold in organisations, ICD effectively acts to make certain organisations easier targets than others for migrant groups which already have co-ethnics employed there. ECD, on the other hand, effectively acts as a form of positive discrimination, constantly counter-acting the tendency of social networks to focus on 'easy' targets (i.e. where the concentration of co-ethnics are high) and forcing them to wait for breakthroughs where they are currently under-represented. This is very effective in reducing labour-market segmentation. --- Discussion and conclusions In this article we have used an agent-based modelling approach to investigate the impact of social networks and employer discrimination on the development of ethnic segmentation in low-skill labour markets driven by referral hiring. Our research has illustrated that ethnically homogenous social networks significantly increase ethnic segmentation in the labour market. Indeed, this impact outweighs that of the Immigration, social networks and the emergence of segmentation <unk> 29 <unk> 'conservative' discrimination procedures tested. Furthermore, such homophilic networks also enhance the ability of immigrant groups to gain a foothold in the labour market, particularly under conditions of discrimination. These findings broadly support the results of an earlier agent-based model of referral hiring and ethnic segmentation (Abdou & Gilbert, 2009), as well as Waldinger & Lichter's (2003) own analysis. However, we have also shown that the form of discrimination has a powerful effect on ethnic segmentation in the labour market. This raises a number of questions and implies that it would be highly beneficial for migration scholars and economic geographers to turn their attention to the initiation, establishment and, crucially, the evolution of migrant social networks and their relationship to the development of ethnic niches in the labour market. Additionally, we must remain aware that the form of discrimination can substantially change its population-level impacts. Following the findings of this study, as well as key debates in economic geography and migration studies, we may ask whether the development of ethnically homogenous social networks are the only successful response to the challenges of gaining employment in low-skill labour markets. Furthermore, this model does not contend with how such networks evolve once a niche has been formed -do they become a barrier to further integration as some theorists have suggested (Lancee, 2012;Ooka & Wellman, 2006), or do they mutate and dissipate as migrant groups learn the language and cultural norms of the host, as human capital theory would imply (Duvander, 2001)? Whilst there is no novelty in stressing the point that ethnic social networks have a significant impact on the labour market success and segmentation of immigrant communities, what we have achieved here is to gauge the relative extent of the impact <unk> 30 <unk> of such networks on labour market segmentation and success, both for different initial cohorts of migrants, and also as compared to 'conservative' discrimination. This has been done with reference to a very specific (though relatively common) type of labour market -one that is driven by referral hiring and with few barriers to employment. The relative impact of ethnic social networks, forms of discrimination and group size may alter substantially where the mechanics of the labour market are different. For example, where entry requirements are higher and require either country-specific qualifications, a level of competence in the host language, or both, the relative strength of impact of these factors is likely to change (see, e.g. Tarvid, 2016). However, we have illustrated that migration scholars and economic geographers can meaningfully model non-linear emergent processes in a way that feeds back into key debates in the disciplines and allows us to combine the reproducibility and

Post-war migration to 'western' countries has gone hand-in-hand with the development of ethnically segmented labour markets, particularly in low-skill roles where entry requirements are minimal. Whilst numerous theories have been forwarded as to why such situations occur, it has remained difficult to empirically test the relative impact of the many interacting processes that produce segmentation in the labour market. In this paper, we investigate the processes of ethnic segmentation in low-skilled labour markets, where referral hiring is the norm, with particular reference to the role of ethnically homogenous social networks and forms of discrimination. We employ an agent-based modelling (ABM) approach, adapting key elements from Waldinger & Lichter's (2003) widely cited networked explanation of ethnic labour market segmentation. This ABM approach allows us to provide a different lens on theories of ethnic labour market segmentation, investigating the relative impacts of different causal processes that are difficult to investigate in this way using other social science ~ 2 ~ approaches. The overall results from our model indicate that ethnically homogenous social networks have the effect of increasing the level of ethnic segmentation within a referral-based labour market, but that these networks also help immigrant populations grow and protect them from the negative impacts of employer discrimination. Furthermore, these networks have a greater impact on labour market segmentation than discrimination alone. In conclusion, this sociologically-informed ABM provides important insights into the manner and extent in which changes in social conditions may affect population-level phenomena.

Whilst there is no novelty in stressing the point that ethnic social networks have a significant impact on the labour market success and segmentation of immigrant communities, what we have achieved here is to gauge the relative extent of the impact <unk> 30 <unk> of such networks on labour market segmentation and success, both for different initial cohorts of migrants, and also as compared to 'conservative' discrimination. This has been done with reference to a very specific (though relatively common) type of labour market -one that is driven by referral hiring and with few barriers to employment. The relative impact of ethnic social networks, forms of discrimination and group size may alter substantially where the mechanics of the labour market are different. For example, where entry requirements are higher and require either country-specific qualifications, a level of competence in the host language, or both, the relative strength of impact of these factors is likely to change (see, e.g. Tarvid, 2016). However, we have illustrated that migration scholars and economic geographers can meaningfully model non-linear emergent processes in a way that feeds back into key debates in the disciplines and allows us to combine the reproducibility and generality of quantitative approaches with a commitment to the nuance and locality inherent in qualitative methods. It is our hope that scholars with a specialisation in either approach will find something here to discuss and extend in future research. In case of several applicants with the highest number of recommendations, a random one of these is chosen. Agents without recommendations therefore only have a chance of employment if none of the existing employees have any unemployed friends applying for a job at the same time. 2 The annual turnover rate for restaurants is missing in the case study data. We have assumed a value slightly higher than those for the other industries to account for the fact that the turnover rate in the restaurant sector is notoriously high. 3 The case study data specifies numbers for native-born and foreign-born Latinos separately. For the sake of simplicity (and to avoid negative numbers for 1960), these have been combined for use in the model.

Post-war migration to 'western' countries has gone hand-in-hand with the development of ethnically segmented labour markets, particularly in low-skill roles where entry requirements are minimal. Whilst numerous theories have been forwarded as to why such situations occur, it has remained difficult to empirically test the relative impact of the many interacting processes that produce segmentation in the labour market. In this paper, we investigate the processes of ethnic segmentation in low-skilled labour markets, where referral hiring is the norm, with particular reference to the role of ethnically homogenous social networks and forms of discrimination. We employ an agent-based modelling (ABM) approach, adapting key elements from Waldinger & Lichter's (2003) widely cited networked explanation of ethnic labour market segmentation. This ABM approach allows us to provide a different lens on theories of ethnic labour market segmentation, investigating the relative impacts of different causal processes that are difficult to investigate in this way using other social science ~ 2 ~ approaches. The overall results from our model indicate that ethnically homogenous social networks have the effect of increasing the level of ethnic segmentation within a referral-based labour market, but that these networks also help immigrant populations grow and protect them from the negative impacts of employer discrimination. Furthermore, these networks have a greater impact on labour market segmentation than discrimination alone. In conclusion, this sociologically-informed ABM provides important insights into the manner and extent in which changes in social conditions may affect population-level phenomena.

Introduction Segregation-the residential separation of people based on race, ethnicity, and/or incomeis an enduring characteristic of modern urban life. Dawkins (2004) offers five explanations for why residential segregation occurs, ranging from discrimination in housing markets and other forms of prejudice to residential preferences that differ by race. Charles (2001) notes that racial groups may reside in different areas because of objective differences in economic status, finding that, on average, whites spend more on housing than blacks. Clark and Ware (1997) report decreasing segregation as income increases. Fisher (2003) finds that while overall racial segregation declined from 1970-2000 the relative importance of income/class to residential segregation increased (p. 685). Segregation is a dynamic phenomenon; decade by decade, observed levels of segregation can change substantially (Logan et al. 2004). Segregation is also a complex phenomenon, shaped by economic factors, individual preferences, racial/ethnic discrimination, and institutions. Segregation is fundamentally a spatial phenomenon having to do with the arrangement of people in space. However, the arrangement of people in space is constrained by the configuration of the built environment-particularly the location of housing. Changes to these constraints may affect the degree of residential segregation. Watson (2006) finds that the rate of growth in housing markets affects both income and racial segregation because in faster growing metros housing markets are better able to adapt to evolving residential preferences. Watson's findings are interesting because they suggest that the built environment of cities may effect residential segregation. In this paper we consider how changes to the built environment of cities relate to changes in observed levels of residential segregation. We model the decentralization of housing and the demise of the walking city, a specific structural change in cities that began in the late 19th century and continued throughout the 20th century. Decentralization of housing is the decline of residential density in the city center and the rise of residential neighborhoods at increasing distances from the center (Bruegmann2005, Hall 1996). We exploit insights from a high resolution historical GIS that maps 100 % of the population of Newark, NJ in 1880 (Logan et al. 2010, Spielman andLogan 2012). We use this data to train a Schelling-like agent based model that reproduces the residential patterns in 1880 Newark. We then use the trained model to explore how changes to the configuration of the built environment, mainly the decentralization of housing, influence ethnic and economic residential segregation. --- Background and Literature Review Space and the configuration of people are central themes in the literature on residential segregation (White 1983, Iceland et al 2002). Methods for measuring segregation have been developed that are sensitive not just to the relative mix of people in an area, but also their spatial arrangement (Wong 2005, Lee at al. 2008). These statistics quantify the arrangement and relative mixture of racial and ethnic groups in defined geographic areas and provide essential tools for describing observed residential patterns. These statistics are useful for descriptive purposes but they do not, in and of themselves, provide insight into the causes and consequences of segregation. On the other hand, dynamic models of residential patterns inspired by Schelling (1971Schelling (, 1978) ) explore the role of various social processes in the formation of segregation (Fossett 2006, Crooks 2010). A simple Shelling model (SSM) contains two types of agents, typically labeled with colors (such as black and white). The "landscape" of an SSM is a lattice of equally-sized squares-similar to a checkerboard. Each square, or "cell," contains one agent and neighbors eight other squares, which also contain one agent each. In an SSM an agent (often a proxy for a household) has a preference to live in a cell where a minimum percentage of its neighboring cells are occupied by agents of the same color. Agents in cells without the minimum percentage of similar neighbors are "unsatisfied," and move to a new cell. Allowing agents to iteratively apply this process of observation and movement results in near complete segregation between the two agent types, even when agents have a significant tolerance for the other group (Zhang 2004). Schelling, in his original formulation of the model, adjusted agents' preferences for living near members of the other group. He found, as have many others, that even slight preferences for living near members of the same group at the individual level produce extreme residential separation at the group level. The impact of racial/ethnic preferences on residential outcomes seems clear and recent work has provided substantial insight into the evolution of preferences over time, however the origins and causes of racial preferences are enormously complex and difficult to explain (Clark and Fossett, 2008). The simple Schelling model has been dramatically extended to include real-estate markets, housing quality, realistic racial/ethnic preference schedules, income, status, and more than two racial/ethnic groups (Zhang 2004, Fossett 2006, Bruch and Mare 2006). For example, Zhang 2004 modifies a simple Schelling model to allow for a housing market, including a vacancy rate. Zhang's model reproduces not only the phenomenon of "white flight" from inner city neighborhoods but also mirrors changes in housing price differentials between white and black neighborhoods. Fossett's (2006) SIMSEG program has been used to simulate geographic variations in housing quality and affordability. Bruch and Mare (2005) and Clark (1991) have used survey data in Schelling-like models to gain insight into the drivers of segregation in real populations. As Crooks (2010) notes, it is useful to think about Schelling-like models "on a continuum between abstract demonstrations to real-world applications" (p. 663). Clark and Fossett (2008) note that even abstract models can reproduce the "complex and sometimes subtle segregation patterns" seen in real cities, thus illustrating "how choices play out in the social fabric and lead to segregated residential outcomes" (p. 4109). We do not believe that these types of models can explain the causes of segregation in the real world, but they can illustrate how micro-level behavior can produce (sometimes unexpected) macro-level patterns. The patterns of segregation produced by Schelling models are part of their appeal. Varying degrees of segregation by race, ethnicity, and income have been the norm in industrial and post-industrial cities in the US. However, this was not always the case. Early in the industrial revolution, before the advent of mass transportation, cities were very dense and more diverse. Charles Booth's maps of poverty in London are a striking example-one notices that there are often very poor streets mixed in among better-off neighborhoods (Dyos et al. 1982). Greenberg (1981), using data from the Philadelphia Social History Project, argues that residential segregation in Philadelphia in the late 19th century was shaped by access to employment. Greenberg (1981) claims that structural constraints, such as the need to walk to work, played a larger role in shaping the urban residential landscape than cultural constraints enforcing homophiliy. The ethnic composition of industries and the spatial distribution of employment within the city played a larger role than preference, according to Greenberg. Areas with industries that tended to be dominated by a single ethnic group appeared to be ethnically segregated. However, areas near industries in which employees were ethnically diverse, and areas where multiple types of employment were concentrated, tended to be diverse. An important exception was for blacks-even though they were a relatively small portion of the population, they appear to be spatially isolated from others in ways that cannot be explained by employment. The basic Schelling model is ill suited to capture the complex residential landscape of the late 19th century. Spielman and Logan (2012) study four cities in 1880 (Newark, NJ; Buffalo, NY; Cincinnati, OH; Albany, NY) using address-level maps of the entire population from the Urban Transitions Historical GIS and find that found that the central (core) areas of these cities tends to be more ethnically diverse, but wealthier and more economically homogeneous than other parts of the city. Further from the city centers, neighborhoods tend to be increasingly mono-ethnic. However, when looking at fully mapped household level census enumeration records, Spielman and Logan find a complex relationship between ethnicity and income. Their results show evidence of ethnic mixing-but they find that residential separation by ethnicity in some areas co-existed with ethnically integrated neighborhoods in others (figure 1). The late 19th century was a period of great change for many cities. Zunz (1982) argues that in Detroit, ethnic and economic segregation dramatically increased between 1880 and 1920. The patterns observed by Spielman and Logan (2012) may only represent one phase in that transition. This transition was closely, if not causally, linked to advances in mass transportation (Hall 1996, Warner 1962). In the 19 th century urban transportation options were limited, and those that existed were expensive. In 1850, a round trip transit ticket in Philadelphia cost about 26% of the daily wage for an unskilled worker. By 1880, the same trip cost 14% of daily wages. However, within 10 years of 1880, mass transportation systems expanded greatly. Hershberg (1981) find that per capita mass transit ridership in Philadelphia increased around 70% every decade from 1870-1890. After 1890 the increases accelerated-by 1910 the average Philadelphian made 200 trips on mass transit per year (p. 153). The relative cost of urban transportation has continued to steadily fall: by 1980 a person earning the federally mandated minimum wage could buy a round trip ticket on mass transit in Philadelphia for just 4.4% of their daily earnings (Hershberg 1981, p. 153). The changes in Philadelphia were characteristic of other cities at the time. Due to the advent of affordable mass transportation, people no longer had to live within walking distance of work, and the hyper-dense urban cores of the late 19th century became the "two-part city" described by Warner (1962). A two-part city is characterized by separate spaces for living and working; the separation of home and work implied by the concept of a two-part city requires widespread commuting. The rise of commuting corresponded with a dramatic shift in the physical structure of cities. Bruegmann (2005, p. 19) plots the population density of London as a function of distance from 1800-1950. These plots show a "flattening out" of the population distribution within the city-the central city became less dense (residentially) and the city became more expansive geographically. Bruegmann (2005) shows that in 1840 once one was eight miles from the center of London population densities were essentially rural, while by 1920 those same regions had nearly 15,000 people per square mile. This shift in the built environment within cities may also have important implications for the development of ethnic and income segregation. We model this transition from a dense compact city to a more expansive, lower density form within the framework of a Schellinglike model. The dynamic framework of a Schelling model lends itself well to questions about the evolution of cities, but until recently data limitations have hampered historical inquiry. Gregory and Healey (2007) note that there have been relatively few analytical studies with Historical GIS. We use recent advances in Historical GIS to provide an empirical foundation for a Schelling-like agent based model. --- RESEARCH QUESTIONS Explaining the co-existence of ethnically segregated and integrated neighborhoods is a challenge for Schelling models (Clark 1991). It has been found that Schelling models have a tendency to produce either segregated or integrated outcomes, but not outcomes where only parts of a city are segregated (Laurie and Jaggi 2003). In this paper we first seek to develop an extended, Schelling-like model that is capable of reproducing the type of patterns observed in the late 19th century-diversity in the urban core and segregation in peripheral neighborhoods (Figure 1). We then use the model to explore how the physical structure of cities can affect levels of segregation along ethnic and economic lines. Specifically, we seek to examine how the shift from a compact, dense urban structure characteristic of cities in the late 19th century to the more expansive, lower-density structure of cities in the 20th century affects the development of the social environment. To our knowledge, these are both novel extensions of the Schelling framework. We believe these questions are both of historical interest and contemporary relevance, as some American cities in the past decade have reversed the trend and experienced significant re-densification through infill development in the urban core. --- DATA The data used to create our agents and validate our model were compiled by the North Atlantic Population Project (NAPP) at the Minnesota Population Center and mapped by the Urban Transition Historical GIS Project (UTP) at Brown University (Logan et al 2011;Minnesota Population Center 2008). Historical Geographic Information Systems like the UTP are becoming increasingly common. Gregory and Healy (2007) note that using GIS in historical inquiry allows one to ask questions about pattern and distribution in ways that non-GIS archival sources do not. Despite the potential for historical GIS to provide novel insights into the geographic organization of society and its evolution, Gregory and Healy (2007) also observe that analytical studies of historical data using GIS are rare. While we generally agree with this observation, there are some notable exceptions (such as Dorling et al. 2000, Hershberg 1981and, Gregory 2008). To our knowledge, this is the first use of a historical GIS in the context of a Schelling-like modeling effort. The NAPP/UTP use the 100% digital transcription of the handwritten records compiled by the individual enumerators who conducted the 1880 decennial census. These digital transcriptions were organized by the Church of Latter Day Saints and are made available by the Minnesota Population Center. For 39 major cities, UTP has augmented the NAPP data through the addition of street addresses for all residents, and has mapped those addresses using historical sources. This study is based on residential patterns observed in a subset of four cities described by Spielman and Logan (2012): Albany, NY, Buffalo, NY, Cincinnati, OH, and Newark, NJ. All four cities were among the top 25 largest cities in the United States. However, we train our model using the observed ethnic segregation patterns in Newark, NJ. Newark's selection as the focal city is largely arbitrary-each of the cities exhibited similar types of settlement patterns (Spielman and Logan 2012). These four cities had similar overall ethnic composition. The primary population groups were German immigrants, Irish immigrants, and native-born whites (Yankees); no other single group represented more than 10% of the population. Germans and Irish are defined here as persons who were born in Germany or Ireland, or had at least one parent who was born in Germany or Ireland. Yankees are whites born in the U.S. with U.S.-born parents. Together, these groups comprised about 80% of the population in each city. The other variable that we draw on is socioeconomic status, as determined by the NAPP using the occupation of the highest status member in a household. Occupations were ranked on a scale from 0-100 based on the average education and earnings of persons in each occupation as measured in 1950. The resulting scale, Duncan Socio-Economic Index (SEI), is commonly used as a measure of socioeconomic status. Sobek (1996) compared the average income of men in each of 140 occupations in 1890 to the income of men in those same occupations in 1950. Sobek finds a correlation between the two of.93, concluding that the SEI scale is valid measure of socioeconomic status for the late 19th century. Using these definitions, in 1880 there were 8783 German households in Newark, NJ, 6247 Irish households, and 7517 Yankee households (n=22547). The highest SEI in our dataset was 96 and the lowest was zero. On average Irish had the lowest SEI (25.3), followed by Germans (31.1), and Yankees (39.1); analysis of variance shows the observed differences in income between ethnic groups to be statistically significant. Figure 2 shows the distribution of income (SEI) within each ethnic group. Using these data, we find that residential segregation by ethnicity was moderate globally, but extremely high in some regions of the city. At the street block level, the Wong (2005) spatial dissimilarity index is 0.53 for Irish and German, 0.39 for Irish and Yankee, and 0.54 for German and Yankee (Xu 2011). Nonstationary patterns such as the ones we observe (in figure 1) can be difficult to quantify. The segregation statistics reported above are somewhat misleading, because they represent an overall metropolitan average and ignore significant local variation. --- METHODS Our model, developed in NetLogo 4.1.3, is an extension of the simple Schelling model. We have made three major modifications to the simple Schelling model: 1. We generate a population of agents that mirrors the population of households in Newark, New Jersey in 1880. Agents belong to one of three ethnic groups (German, Irish, or Yankee). The agents are assigned an income (SEI) based upon the observed distribution of incomes in 1880. 2. We parameterize the landscape upon which agents interact. This parameterization distributes housing units across the landscape to create housing density profiles like those observed in late 19th and early 20th centuries. While we have experimented with a number of different shapes of density profiles, here we focus our attention on the simplest case, in which we distribute housing units using various Gaussian distributions. 3. We develop a utility function for the agents that allows us to reproduce the observed geographic relationships between ethnic composition and incomes. Our work falls somewhere in the middle of the continuum between pedagogic and realistic Schelling models. The characteristics of our agents are empirically derived, but the mechanisms they use to select residential locations are too simple and stochastic to reflect reality. Fossett (2006) and Zhang (2004) have developed sophisticated residential choice algorithms for Schelling models that account for housing quality, price, and demand. Our model was reliably able to reproduce the patterns similar to those we observed in 19th century Newark while considering only the ethnic and economic preferences of the agents. The model emphasizes parsimony, rather than trying to capture all variables that may have an effect on observed residential patterns. The only variables included in the model are directly measured by the 1880 census. The model does not include a real estate markets or complex housing search procedures as such mechanisms are beyond the scope of the current investigation. We used this model to conduct two distinct experiments. The first is designed to examine how the model responded to variations in parameters, and to determine combinations of parameters that produce residential patterns in the model similar to those observed in the empirical 19th century data. Once we identify parameter regimes that reproduce the empirical patterns, the second experiment vary the housing density profiles of the landscape on which the agents interact. The second experiment is designed to mirror changes to urban environments that occurred in the early 20th century. It seeks to understand if the evolution of the built environment, holding preferences for ethnicity and income constant, results in changes in observed patterns of segregation. This second experiment is a novel extension to the Schelling framework that explores the role of physical environmental constraints on segregation. Previous work by Fossett (2006) has explored the role of spatial variations in housing quality and price on segregation, but not the role of urban from or housing unit density. --- Creating Agents In generating agents for our model, we seek to preserve both the relative proportion of each ethnic group in the population and the income distribution within each group. Mirroring the relative proportion of households in Newark, the population of agents in our model is 39% German, 28% Irish, and 33% Yankee. Figure 2 shows the SEI distribution used for each ethnic group, generated empirically from the Newark data. While we use SEI as a proxy for income, it is important to realize that SEI is a quantification of a qualitative variable. The census recorded people's professions, so common generic answers such as "laborer" result in spikes in figure 2 (such as the spike around 18 in all three plots). It was not possible to find statistical descriptors of these irregular distributions. Instead, for each ethnic group we divide the empirical income distribution into sixteen discrete "bins." Each agent has two characteristics that remain fixed throughout an entire run of the model: 1. ETH: Ethnicity which can take the value "German," "Irish," or "Yankee" 2. SEI: Income, assigned based upon ethnicity-specific income distribution. Where ETH i and SEI i refer to the ethnicity and income of the i th agent. --- Creating Landscapes Crooks (2010) notes that the landscapes used in Schelling models tend to be flat planes on which a single agent inhabits each cell. Our work partially overcomes these limitations; we generate landscapes such that housing units are distributed among the cells according to a bivariate, discretized Gaussian distribution with mean of zero and a user-specified variance. This means that cells can house a variable number of agents, and cells near the "city center" can house more agents than cells on the periphery. We use a single variance parameter, which we refer to as the housing density distribution parameter, resulting in an isotropic distance decay of housing units from the center. For example, when the housing distance decay parameter is fixed at 7.5 roughly 95% of all housing units are within 15 cells of the center of the landscape, when it is fixed at 4.5, 95% of all units are within 9 cells of the center. These landscapes roughly reflect population density gradients of industrial cities during the late 19 th to early 20 th centuries, a period during which the rise of mass transportation fundamentally restructured cities (Hall 1996). Bruggemann (2005) shows that London's density profile evolved from, what in our model would be a very small setting of the housing distance decay parameter in 1841 to a what would be a large housing distance decay parameter in 1921. During this period the population density in London was essentially log linear with respect distance from the city center, this relationship is approximated by the housing density distribution parameter. By 1950 London's population distribution had characteristics of a log-normal density profile. We have explored more complex functional forms, including "donut" cities with an uninhabited core and anisotropic surfaces. We limit our discussion to the symmetrical Gaussian case, because it provides a reasonable and parsimonious description of cities during our study period. The landscape is generated such that it has a vacancy rate of approximately 3-5%, that is, there are always 3-5% more housing units than there are agents. Changes to the housing density distribution parameter alter the spatial distribution of housing units but the number agents, and therefore the number of housing units, remains fixed. We run our model on a landscape of 33<unk>33 cells. Depending on the housing density distribution parameter selected, some number of these cells have a housing capacity of zero, meaning that they cannot accommodate agents and are not active in the model. --- Utility Function Agent behavior is controlled by utility, which is determined by a specialized utility function and measured along a 0-1 scale. The utility function contains an ethnic term (left term in equation 1 below) and an economic term (right term in equation 1 below). Each term is weighted by agent-specific weight w i, calculated as, where SEI max is the maximum SEI in the city of Newark (96) and ME is a model parameter called Money Exponent (ME). For any agent i, w i will be bounded between 0 and 1. Agents with higher SEI values will have values of w i closer to 1, while agents with lower SEI values will have values of w i closer to 0. If the Money Exponent (ME) is set to 1 the weight (w) increases as a linear function of SEI. This weight term is described in figure 3, which plots the weight as a function of SEI for several values of ME. An agent's utility at a given location in the landscape is based upon both its own characteristics and the characteristics of the other agents within its field of view (equation 1). The field of view is determined by a model parameter called Agent Vision and denoted r. This vision parameter encodes the amount of information about the city available to agents in the calculation of utility. Vision has been viewed as a way to represent a host of factors that limit (or enhance) an agent's knowledge of its environs, such as real estate agents (Laurie and Jaggi 2003). To calculate the utility of its current location, an agent identifies all agents within r units of itself. We denote this set N i (r). For each agent j in N i (r) we evaluate it ethnicity relative to the ith agent-<unk> ETH i,ETH j takes a value of 1 if agent i and j have the same ethnicity otherwise 0. Hence, the ethnic term of the utility function is simply the proportion of all agents near agent i that are of the same ethnicity. The right term in the utility function measures the relative affluence of the agents within r units of the focal agent (i). We denote the median SEI within r units of the i th agent using. (1) The ethnic term (left) is multiplied by 1-w, and the economic term (right) is multiplied by w. This causes lower-income agents (w closer to 0) to place relatively more weight on the ethnic term than the economic term, and higher-income agents (w closer to 1) to place relatively more weight on the economic term than the ethnic term. In both the historic and modern contexts, it is common for ethnic segregation to decline as a function of income (Clark andWare 1997, Spielman andLogan 2012). This aspect of our utility function captures this phenomenon. The Money Exponent is central to the calculation of utility, as it determines how much weight most agents place on the ethnic term and the economic term. An agent near the upper end of the income distribution will always have a w near 1, while an agent with an SEI of 0 will always have a w of 0. The large majority of agents that fall somewhere in between on the SEI scale can place quite different weights on the ethnic term and economic term, depending on the value of ME (Figure 3). --- MODEL EXECUTION At the initialization of each model run, all agents are placed onto the landscape into a random cell with an unoccupied housing unit. This random initial condition creates an integrated city. The model then proceeds iteratively, with every agent evaluating its utility during each iteration. Agents calculate their utility using the utility function described above, and compare the result to the utility threshold, a user-defined parameter on a 0-1 scale. If an agent's utility is below the threshold, it moves to a new random cell with an unoccupied housing unit. This process continues until one of three stopping criteria is met. The first is based on the agent utilities: if 98% of the agents are satisfied (i.e. their utility exceeds the utility threshold), the model run terminates. The second stopping criterion is met if the number of unsatisfied agents has not changed by at least 10% in the last 20 iterations. This condition implies that the model has approached a steady state with fewer than 98% of the agents satisfied. The final stopping criterion is based on time: if the model has not met one of the other stopping criteria after 200 iterations, the run terminates. At model termination, we output a series of statistics describing the aggregate urban pattern. We calculate pair-wise spatial and non-spatial segregation indices for all ethnic groups, using both the non-spatial dissimilarity index (D) and Wong's (2005) spatial measure of dissimilarity. We also output the correlation between population density and income, the average utility of the agents, the percent of agents that have a utility exceeding the utility threshold, and the number of iterations required for the run. Figure 4 shows output from three model runs and some of the associated statistical descriptors. Each model is run with the same set of parameters but using different "landscapes," or spatial distributions of housing units. Cells are colored according to the relative ethnic mix of the agents located there, with red for Germans, green for Irish, and blue for Yankees. Cells that are solid red, green, or blue represent ethnically homogenous areas, while cells of intermediate colors represent significant ethnic mixing. The brightness of each cell corresponds to the population density. The white center region in example 1 is a hyper-dense urban core. In the first example, 95% of all housing units are within 6 cells of the city center. The segregation indices observed tend to increase as the landscape becomes less dense and more spread out. In all three examples, some degree of ethnic mixing can be observed in the city center, while ethnically homogeneous neighborhoods tend to form at the periphery. The examples below are just three of many thousands of model runs conducted. --- EXPERIMENTAL DESIGN --- EXPERIMENT 1: Full parameter sweep-We use a simple experimental design to identify combinations of model parameters that produced similar residential patterns to those observed in the empirical 1880 data. We specify a range of values for the money exponent, utility threshold, housing density distribution, and agent vision parameters and take their cross product, creating a four-dimensional parameter space. The Money Exponent parameter takes values between 0.2 and 1.95 incremented by.35. The utility threshold takes values between.2 and.6 incremented by.1, and the agent vision parameter takes values between 1 and 5 incremented by 1. Housing density distribution parameter takes values from 4.5 to 10.5 incrementing by 1. For each combination of parameters we conducted 8 independent model runs, resulting in 10,080 independent model runs. We describe the results of these runs in the following section. --- EXPERIMENT 2: Varying the built environment-We conduct a second, smaller experimental run in which we use a set of model parameters that was capable of reproducing the type of residential segregation patterns observed in Newark, NJ. The purpose of this experiment is to explore how the evolution of urban form that occurred between the late 19th and early 20th century could affect residential segregation patterns, holding preferences constant. We hold the entire set of parameters constant, except the housing density distribution parameter, which we vary from 2 to 15 units ("varying the built environment"). When the housing density distribution parameter is small, housing is concentrated in the core (like a walking city); when it is large, the core is less dense and housing is spread further from the center (as in a transit-oriented city). --- MODEL VALIDATION Our model places an empirically-based population of agents onto an abstract surface, which makes it difficult to validate. This population matches the ethnic and economic composition of Newark, New Jersey in 1880. Our population includes only the three main ethnic groups, which together constitute roughly 80% of the city's population. However, these agents interact on an abstract space that does not resemble Newark's geography. The model is validated based upon its ability to reproduce residential settlement patterns with statistical similarities to the NAPP data and the patterns observed by Spielman and Logan (2012). The parameter regime that produced residential segregation statistics closest to those observed in 1880 was Money Exponent=0.55, the Utility Threshold of 0.4 or 0.6 and agent vision between 3 and 5. This parameter regime produced segregation statistics for each of the pairwise groups that were within 0.01 of the observed value. We refer to this set of parameters as the "fitted model" and used them to conduct the second experiment. We also examined the relationship between SEI and proximity to the urban core. In 1880, as one moved out from the city center, SEI tended to decline. At the end of each model run, we recorded the correlation between the population density of each cell and the SEI of agents residing in that cell. When using landscapes with a dense core (i.e. a low value of housing distance decay parameter), we consistently found positive correlations, suggesting that the wealthier agents tended to cluster together in the city center. However, there was some variability in this correlation-it ranged from a high of 0.8 to a low of -0.2 with a mean of 0.4. The correlation between population density and SEI is sensitive to the landscape, a phenomenon we hope to explore in the future. --- RESULTS --- Experiment 1 Across the 10,000+ model runs in the first experiment, there is high degree of correlation among the pairwise spatial and non-spatial dissimilarity statistics between the different groups (figures 4 and 5). In our data, we saw the lowest segregation among Irish and Yankee groups and higher levels of segregation between Germans and Irish, and Germans and Yankees. In Newark in 1880 we notice especially strong segregation between low income Irish and low income Germans. In the real world, this phenomenon is probably due to a complex set of social relations, including the linguistic isolation of German-speaking immigrants. German immigrant communities in the US in the late nineteenth century formed self-contained neighborhoods in which the everyday language was German (Zunz, 1982). Our model does not include language, however it does include group-specific income distributions. The Irish and German populations have relatively large segments in the lower end of the SEI distribution. These population segments will place more emphasis on ethnicity, resulting in ethnic segregation. This may create the lower correlations between the Irish-Yankee segregation statistics and the German-Irish segregation statistics observed in Table 2. Given the high degrees of correlation we observe between the different measures of segregation, we focus our attention on a single outcome-the German-Irish spatial dissimilarity statistic. Figure 5 shows how German-Irish segregation (vertical axis) responds to changes in the various model parameters (horizontal axes). They include the data from the 10,000+ model runs in the first experiment. These plots should be interpreted with caution; for example, the plot showing the utility threshold (lower right panel of figure 5) shows the level of segregation when the threshold is set to 0.2, 0.3, 0.4, 0.5, and 0.6. For each of those 5 levels the other parameters in the model vary-the plot does not statistically isolate the effect the parameter (i.e. it is not an effect plot). The lines in each of the four panels in figure 5 are locally weighted scatterplot smoothing fits (LOWESS). In general, these plots show that there is substantial heterogeneity in the level of segregation produced when one parameter is held fixed and the others are allowed to vary within the experimental parameter space. One notices that increases in the housing density distribution, utility threshold, and the money exponent parameters all, on average, increase the level of segregation in the resulting model. However, there seems to be a point tipping point in the utility threshold parameter, beyond which further increases result in decreases in segregation. Through further experimentation we found that in a wide variety of parameter regimes including a utility threshold value above.55, the agents in the model failed to settle into stable patterns. Under these circumstances, many agents were unable to find any location that provided satisfactory levels of utility; in these instances the resulting segregation statistics are unusually low (below 0.2). Increases in the agent vision parameter, on average, resulted in decreases in the level of segregation. However, few of the data points are near the LOWESS line-most are clustered in outcomes with a high or a low level of segregation. These clusters of high and low values can be somewhat explained through the interaction plots in Figure 14. These interaction plots show the average level of German-Irish segregation for every pairwise combination of model parameters. Horizontal and vertical axes take their minima and maxima from the parameters ranges used in experiment 1 (see experimental design section). In this graphic, the color values of each pixel are determined by the average level of segregation for all model runs containing that combination of parameters. These plots provide a visual summary of how the model responds to variations in

To what degree does the built environment of cities shape the social environment? In this paper we use a Schelling-like agent based model to consider how changes to the built environment of cities relate to changes in residential segregation by income and ethnicity. To develop this model we exploit insights from a high resolution historical GIS which maps 100% of the population of Newark, NJ in 1880. Newark in 1880 had a complex social landscape characterized by areas of significant social and economic segregation and areas of relative integration. We develop a Schelling model capable of reproducing these residential patterns. We use this model to explore the decentralization of housing, a specific phenomenon associated with the demise of the walking city in the late 19th century. Holding agent preferences constant, but allowing the landscape of the Schelling model to evolve in ways that reflect historical changes to the built environment produces changes to the social landscape that are also consistent with history. Our work suggests that changes in residential segregation do not necessarily imply changes to individual attitudes and preferences. Changes in residential segregation can be generated by changes to the built environment, specifically the geographic distribution of housing.

a wide variety of parameter regimes including a utility threshold value above.55, the agents in the model failed to settle into stable patterns. Under these circumstances, many agents were unable to find any location that provided satisfactory levels of utility; in these instances the resulting segregation statistics are unusually low (below 0.2). Increases in the agent vision parameter, on average, resulted in decreases in the level of segregation. However, few of the data points are near the LOWESS line-most are clustered in outcomes with a high or a low level of segregation. These clusters of high and low values can be somewhat explained through the interaction plots in Figure 14. These interaction plots show the average level of German-Irish segregation for every pairwise combination of model parameters. Horizontal and vertical axes take their minima and maxima from the parameters ranges used in experiment 1 (see experimental design section). In this graphic, the color values of each pixel are determined by the average level of segregation for all model runs containing that combination of parameters. These plots provide a visual summary of how the model responds to variations in parameters. They are not to be confused with output from the model. The right-most plot on the bottom row of figure 6 shows the interaction of agent vision and housing density distribution parameters. The interaction between the two parameters is nonlinear: very high and very low values of agent vision are, for most values of the housing density parameter, associated with low values of segregation. However, when in the middle of the values range for the agent vision parameter of the housing density parameter segregation levels tend to be high. Generally, these interaction plots suggest complex and non-linear interactions among parameters-for example, the interaction between money exponent and utility threshold is difficult to explain. Because of the complexity and variability in our analysis of the full parameters space, we focus on a narrow set of parameters in experiment 2. These parameters, the "fitted model," were selected from the results of the first experiment because they closely resemble the patterns we observe in the late 19th century-a relatively ethnically diverse and high-income central city surrounded by ethnically segregated, lowerincome neighborhoods. --- Experiment 2 In the second experiment, we held all parameters constant at their "fitted model" values, except for the housing density distribution, which we allowed to vary. We find that changes to the physical structure of the city, holding all other model parameters constant, produces large changes in the average level of segregation (figure 7). Compact, dense city forms tend to result in lower levels of ethnic and income segregation than less dense, more expansive forms. We notice this result in both the first and second experiments, though the effect is more pronounced in the second experiment with the fitted model. For higher values of the housing density distribution parameter, the difference in housing capacity between the city center and the periphery is less pronounced, and housing units are more evenly distributed across the landscape. This causes the configuration of the model to become more like the traditional Schelling framework, where housing units are distributed uniformly on a plane. Under these conditions, the model output moves closer to the traditional Schelling result of strong ethnic segregation, despite the fact that the higher income agents place little utility on ethnic homogeneity. The more expansive space of the model allows agents to sort themselves by both ethnicity and income, resulting in higher segregation along both ethnic and economic lines (figures 7 and 8). --- Conclusion While not entirely surprising, the second experiment provides an important and intriguing result. In the social sciences it has long been held that the physical arrangement of people in space both reflects and shapes attitudes and social processes (Abbot 1997). In the literature on residential segregation these attitudes are often understood to be some type of preference -a preference for a neighborhood with a certain ethnic composition, housing stock, and/or economic characteristics (Dawkins 2004). The social processes relate to the economics of housing markets, discrimination, and differential access to information. However, we find that by altering the built environment of a city, while holding all social factors constant, we are able to produce substantial changes in segregation outcomes. A city's physical structure plays an important role in the manifestation of social attitudes and preferences. We develop this insight by allowing a fixed set of attitudes and a fixed process to play out on a varying landscape. Our changes to the built environment mirror historical changes that happened to many cities in the late 19th and early 20th centuries. During this period, segregation by ethnicity, race, and income increased substantially in some cities (Zunz 1982). Our results mirror these increases. However, we are not suggesting that the changes to the social landscape of cities around the turn of the last century were necessarily due to changes in the built environment. It is a fact that the two occurred contemporaneously, but a Schelling-like agent based model cannot establish causation. Our model establishes a plausible and novel contributor to observed metropolitan segregation patterns. We find that the built and social environments of cities co-evolve. In discussions of segregation this co-evolution does not receive sufficient attention. How significant is the effect of the built environment relative to other factors such as preferences, economic constraints, and discrimination? Do preferences evolve more or less rapidly that the built environment? We are unable to answer these questions, but clearly, they affect how one interprets changes in residential segregation. The changes in residential segregation produced by our model are quite pronounced. We believe that these changes raise epistemological questions about the inferences one can reliably draw from changes in socio-spatial patterns. Our work suggests that changes in the observed level of segregation do not necessarily require changes to attitudes, preferences, or institutions. Over the past 20 years, segregation has generally declined in the US. Concurrently, there have been significant increases in residential density in many inner cities (resulting in changes to the spatial distribution of housing units). Our model raises a question-might some portion of the recent declines in residential segregation be due to changes in the physical structure of cities? We believe this question warrants further investigation. Residential Neighborhood Types in Newark, NJ (1880) Income by Ethnic Group Newark, NJ 1880 Experiment 2: Urban Form and Ethnic Segregation Experiment 2: Urban Form and Income Segregation --- Examples of Model Output --- Model Parameter Descriptions --- Model Parameter Name Description Agent Vision (r) This parameter determines how far agents will "look" (when compiling the set of other nearby agents) while calculating their utility. Money Exponent (ME) Determines how an agent's own income affects the relative weights that agent places on the ethnicity and income level of other nearby agents. Utility Threshold Agents are "satisfied" with their current location if their utility (see eq. 1) is equal to or greater than the utility threshold. Housing Density Distribution Determines the housing density profile of the city. Higher values lead to a concentration of housing in the center. Lower values lead to a more even distribution of housing units. --- Supplementary Material Refer to Web version on PubMed Central for supplementary material.

To what degree does the built environment of cities shape the social environment? In this paper we use a Schelling-like agent based model to consider how changes to the built environment of cities relate to changes in residential segregation by income and ethnicity. To develop this model we exploit insights from a high resolution historical GIS which maps 100% of the population of Newark, NJ in 1880. Newark in 1880 had a complex social landscape characterized by areas of significant social and economic segregation and areas of relative integration. We develop a Schelling model capable of reproducing these residential patterns. We use this model to explore the decentralization of housing, a specific phenomenon associated with the demise of the walking city in the late 19th century. Holding agent preferences constant, but allowing the landscape of the Schelling model to evolve in ways that reflect historical changes to the built environment produces changes to the social landscape that are also consistent with history. Our work suggests that changes in residential segregation do not necessarily imply changes to individual attitudes and preferences. Changes in residential segregation can be generated by changes to the built environment, specifically the geographic distribution of housing.

Background By June 1, 2020, US cases of the novel coronavirus (COVID-19) had surpassed 1.7 million, including more than 104,000 cumulative deaths [1]. State and local officials have actively combatted the disease with prevention tactics, including stayat-home orders, business restrictions, face mask recommendations, and contact-tracing procedures, that have surely reduced the spread of COVID-19. Nevertheless, COVID-19 has continued to spread in many areas of the USA. Data has revealed that African-Americans and other racial and ethnic minorities are suffering the greatest burden of COVID-19 outcomes. For instance, African-American people in Wisconsin have comprised 20.5% of all COVID-19 deaths while only making up 6.3% of the total population, according to a study from the Kaiser Family Foundation that used a Racial Disproportionality Index (RDI) [2,3]. The study findings indicated that African-American people in Wisconsin died at 3.3 times the frequency expected (as of August 17, 2020), given their total representation in the state. Meanwhile, the COVID Tracking Project has reported the national COVID-19-related mortality rate for African-Americans was 2.4 times higher than that of Caucasian Americans [3]. Similarly, a report from the Commonwealth Fund detailed racial disparities across the nation based on county-level data, showing that counties with higher African-American populations experienced greater case, mortality, and progression rates of the disease than counties with lower African-American populations [4]. The specific causes of these disproportionate COVID-19 mortality rates among African-American people are not clear, however they are likely related to persistent and historical disparities in access to care, quality treatment, and social determinants of health [5]. Recent literature has hypothesized several factors that may account for racial differences in COVID-19 mortality, including (1) disparities in environmental factors, such as increased air pollution in areas with high African-American populations, which could exacerbate lung complications [6]; (2) increased rates of comorbidities that negatively affect COVID-19 outcomes [1]; (3) differences in employment patterns and related exposure [7]; (4) disparities in access to care, including lack of insurance and geographic maldistribution of healthcare services [7,8]; and (5) structural racism in the healthcare system resulting in biased, limited care [9]. The COVID-19 pandemic has exposed environmental health factors that contribute to increased mortality among African-American populations. Brandt et al. (2020) reported that increased air pollution impacts COVID-19 mortality both directly, by decreasing lung capacity during COVID-19 infection, and indirectly, by increasing incidence of air pollutionrelated diseases that increase COVID-19 mortality [6]. Due to structural racism in the USA, air pollution has historically been higher in African-American communities [6]. Recent literature has suggested that prolonged exposure to air pollution increases the likelihood of associated diseases, increasing COVID-19 severity and potentially accounting for some of the disparity in African-American COVID mortality rates [6]. Moreover, African-Americans are more likely to have health conditions that increase COVID-19 disease severity, such as obesity and asthma [1]. According to the CDC [10], nonHispanic African-American populations have high rates of obesity and are more likely to experience conditions that contribute to obesity, including unemployment, food insecurity, access to less nutritious foods, and advertising for these poorquality foods. Similarly, the risk of mortality due to asthma is three times higher among African-Americans than Caucasian Americans [11]. This high rate of asthma has been linked to segregation policies like redlining, resulting in a disproportionate number of African-American people living in older houses or communities with high lead exposure [12], which is believed to increase asthma severity and worsen COVID-19 symptoms [13]. Additionally, various occupational disparities also increase the probability that minorities will be exposed to COVID-19. For example, African-Americans and Hispanics are two times more likely to work in animal slaughtering and processing industries where COVID-19 outbreaks have taken place [14]. In addition, compared to Caucasian populations, African-Americans are 4% more likely to have occupations that require working in close proximity to others [14], thus increasing their risk of infection. Higher employment rates, especially in close proximity "essential-worker" positions like healthcare, transportation, and other service industry jobs, correspond with increased COVID-19 infection rates, and these jobs are disproportionately held by African-Americans [7]. Disparities in both access to and delivery of quality healthcare for African-Americans compound these factors. Examples of disparities related to access to care and treatment include lower insurance levels among African-Americans, Hispanics, and a number of Asian groups compared to Caucasians [15]. Lower county-level insurance coverage rates have been associated with both COVID infection rates and African-American population proportion, and researchers have hypothesized that the lack of insurance coverage for African-Americans partly accounts for the higher COVID mortality among this demographic group [7]. A recent study in Georgia found that counties with the highest COVID mortality rates had both the highest African-American populations and the lowest proportions of intensive care unit beds and primary care providers [8]. Additionally, structural racism and bias in the medical field have emerged as a further major concern potentially impacting African-American mortality. Researchers have pointed to both structurally inherent bias leading to poorer care and barriers for access to care leading to greater likelihood of delayed care and complications for African-Americans [9]. For example, African-American populations are less likely to receive appropriate treatment for myocardial infarctions, according to one study [16]. Therefore, we wanted to learn more about the relationship between COVID-19 and race. Only 26 US states were reporting COVID-19 mortality rates by race as of April 17, 2020 [17]. Among these states was Pennsylvania (PA), which, as of August 31, 2020, had around 7673 total COVID-19 deaths [18]. According to 2018 data, PA is about 80.1% White, 11.2% African-American, 3.5% Asian, and about 5.1% reporting as other races [19]. However, the race distribution is not equal throughout the state; Philadelphia county is 42.3% African-American while the remaining counties are 21.4% African-American or lower [20]. This difference in demographic diversity is what led us to study PA in particular. However, while PA has reported deaths by race overall, it has not reported them on the county level [18]. To analyze relationships between mortality rates and risk factors in PA and throughout the USA, we visualized and compared the geographic distribution of COVID-19 death rates and overall African-American populations. The purpose of our ecological study was to explore correlations between COVID-19 mortality and the distribution of the African-American population by US states and within the state of PA. Additionally, we conducted statistical modeling to learn whether these correlations persisted after controlling for other potential confounders, including demographics, healthcare access, behavioral risk factors, and chronic diseases. --- Methods A base map of the contiguous USA was downloaded from the US Census TIGER/LINE databank [21]. COVID-19 death rates for US states and PA counties were downloaded from PolicyMap. The source data for COVID-19 rates was from The New York Times, based on reports from state and local health agencies on May 27, 2020 [22]. The source of the estimated percentages of African-American populations was the US Census American Communities Survey (2014-2018) and was downloaded via PolicyMap. Potential confounders explored included the following: median household income, hospital beds per population, primary care providers per population, self-reported primary care access, obesity rate, smoking rate, high blood pressure rate (HBP), chronic obstructive pulmonary disease (COPD)/emphysema/chronic bronchitis rate, and diabetes rate. Statistics on primary care provider rates and hospital bed rates were derived from the Human Resources and Services Administration (HRSA) via PolicyMap (2016). Self-reported percentages of having a personal provider, smoking cigarettes regularly (adults), having COPD, emphysema, or chronic bronchitis, having diabetes, having HBP, and being obese (BMI <unk>30) were sourced from the CDC Behavioral Risk Factor Surveillance System (BRFSS) via PolicyMap (2017). This information was mapped using ESRI ArcGIS version 10.3.1. We imported data from PolicyMap into Microsoft Excel, edited to ensure that geographic linkage variables corresponded between programs, and joined tabular Excel data to basemaps in ArcGIS. We created choropleth maps of COVID-19 mortality rates across the USA (state-level) and in PA (county-level), with category cutoff values created using natural quantile breaks. We added estimated percentages of African-American populations to each map as graduated point data, with larger points representing larger percentage of African-Americans within the population. We then used SPSS to explore statistical correlations between variables. We log transformed (Log base 10) COVID-19 mortality to use as our main dependent variable, due to the skewed nature of the mortality data [23]. We excluded US territories from the state analysis, and all PA counties with COVID-19 mortality rate of zero from the county analysis (including Clearfield, Potter, Crawford, Somerset, Warren, Elk, Venango, Jefferson, Greene, Sullivan, Forest, Cameron, and Montour counties). Using both the state and county data, respectively, we performed bivariate analyses between COVID-19 mortality and all potential risk factors, including percentage African-American, using Pearson's correlations. The bivariate analyses were used to determine which variables to include in multivariable linear regression models to predict COVID-19 mortality rates. We used linear regression models to control for confounding variables (risk factors) that were correlated with COVID-19 mortality in the bivariate analyses. For all statistical analyses, significance was determined based on a p value <unk> 0.05. --- Main Findings As of May 27, 2020, COVID-19 mortality rates were higher in the eastern region of the USA compared to central and western regions, with the highest rates (around 53.58-149.55 deaths per 100,000 population) in the northeast (Fig. 1). The percentages of African-American populations (US Census, 2014-2018) showed similar descriptive patterns as COVID-19 mortality rates, increasing from the west towards the northeast. All northeastern states with mortality rates within the highest category (53.58-149.55 deaths per 100,000 population) also had the highest percentage of African-American populations (22.12-46.94%), with the exception of Connecticut, Rhode Island, and Massachusetts. In these states, African-American populations ranged between 6.55-10.56%. Many states with low percentages of African-Americans also had low death rates. For instance, in California, 5.79% of the population was African-American, and the mortality rate was 10.12 deaths per 100,000 population, and in Montana, 0.44% of the population was African-American, and 1.63 deaths occurred per 100,000 population. Looking specifically at PA, COVID-19 mortality rates were similarly high in the eastern part of the state (Fig. 2). Several counties in this region, such as Philadelphia and Monroe, had high COVID-19 death rates ranging from 52.95-95.12 deaths per 100,000 population. These eastern counties also had comparably high African-American population sizes between 3.3 and 4.8%. Counties with the lowest percentages of African-American populations tend to be in north-central or northwest PA, with the majority of counties also having low COVID-19 mortality rates. For example, the population of Elk County was 0.55% African-American and the COVID-19 mortality was 0, and the population of Tioga County was 0.66% African-American with a COVID-19 mortality rate of 4.85 per 100,000. --- Statistical Analyses For US states and for counties within PA, we used Pearson's correlations (r) to identify significant relationships between log-transformed COVID-19 mortality rates and the following variables: percent African-American population, median household income, hospital beds per population, primary care providers per population, self-reported primary care access, obesity rate, smoking rate, HBP rate, COPD/emphysema/ chronic bronchitis rate, and diabetes rate. Variables that were significantly related to COVID-19 mortality were included in a multivariable linear regression model to remove the influence of confounding. For US states, COVID-19 mortality was only significantly correlated with percent African-American population (r = 0.445, n = 51, p <unk> 0.001). As other variables were not significantly related to COVID-19 mortality, we chose to run a simple linear regression model to quantify the relationship between these variables (Table 1). To interpret the linear regression model, we removed the log-transformation, and found that for every one-percent increase in the African-American population, there was a 2.12% increase in COVID-19 mortality among US states and the District of Columbia. For PA counties, our bivariate Pearson correlation analysis identified positive correlations between COVID-19 mortality and percent African-American (r = 0.400, n = 54, p = 0.003) and between COVID-19 mortality and income (r = 0.492, n = 54, p <unk> 0.001). Conversely, we observed negative correlations between COVID-19 mortality and COPD (r = -0.448, n = 54, p = 0.001); and COVID-19 mortality and smoking (r = -0.427, n = 54, p = 0.001). Our multivariable linear regression model to predict COVID-19 mortality (Table 2) showed a positive linear association between COVID-19 mortality and percent African-American among PA counties, after controlling for median household income, percent smoking, and percent diagnosed with COPD. After controlling for potential confounding variables that were correlated with log transformed COVID-19 mortality in our multivariable regression model, we found that the relationship between % African-American and log-transformed COVID-19 mortality rates persisted. For every unit increase in % African-American, there was a 0.035 unit increase in logtransformed COVID-19 mortality rate (t = 2.870, p = 0.006), controlling for median household income, percentage smoking, and percentage diagnosed with COPD. To interpret the linear regression model, we removed the log-transformation, and found that for every 1% increase in the African-American population, there was a 3.56% increase in COVID-19 mortality in PA counties. --- Discussion This article visually and statistically analyzes US state-level and PA county-level COVID-19 mortality rates and the distribution of African-American populations. Variables associated with COVID-19 mortality were included in a multivariable linear regression model to remove potential confounding. Due to the ongoing nature of this pandemic, these findings should be regarded as an interim appraisal. Still, worth noting are various geographical trends on both levels. As of May 27, 2020, COVID-19 mortality rates were higher in US states and PA counties in the east of both regions. This comes as no surprise given that New York City was the "epicenter" for the pandemic [24] and that eastern PA counties have higher population density than the rest of the state [25]. Likewise, African-American population sizes appeared highest in these locations. Pearson correlations were used to identify positive associations between percent African-American population and log-transformed COVID-19 mortality rates for the USA and PA. However, it is interesting that, on the national level, COVID-19 mortality was only significantly correlated with percent African-American population and not other factors such as diabetes or COPD rate. In PA, COVID-19 mortality was significantly correlated with percent African-American population but also median household income, percent smoking, and COPD rate. When the above risk factors were controlled for, the association between percent African-American and COVID-19 mortality persisted, demonstrating that the association between African-American population and COVID-19 mortality cannot be "explained away" by any of these potential confounding variables at the group level. As it is unlikely that there is a genetic explanation for higher mortality rates among African-American populations [26], our findings suggest that this relationship may be due to social and environmental factors. Examples of such factors include racial bias in healthcare resulting in lower access to quality and affordable healthcare, both of which drive health inequities apparent in this pandemic [27]. This inequity may be due to delayed access in care because of lower rates in health insurance coverage, according to Poteat et al. [28]. Compared to Caucasian Americans, African-Americans are 1.5 times more likely to either be uninsured or underinsured [28]. Also, in some states, African-Americans with cough and fever symptoms were found to be less likely to receive a COVID-19 test than their Caucasian counterparts [29]. The findings of this investigation stress the importance of equity in addition to equality. Whereas equality focuses on distributing similar resources to all people, equity emphasizes customized distributions that account for disparities between groups [30]. Research suggests that inequitable access to affordable and quality healthcare partially explains unequal COVID-19 mortality rates between races in the USA. It is therefore vital that public health efforts work to eliminate these inequities and provide a customized prevention approach for African-American communities, in order to mitigate unequal health outcomes during the current pandemic, and prevent them for the next. Our findings also indicate the importance of tracking data on race as it relates to COVID-19. States across the nation should offer more data on race and COVID-19 mortalities, including county-level data, so that public health officials can have adequate surveillance data to identify and, ultimately, address the causes of disparities by race. Similar calls to action are taking place across the country [31], as this step is crucial to fully understand the extent of this disease burden. Additionally, our findings of COVID-19 mortality trends and associations among African-Americans can inform policy on where to designate resources. These findings also emphasize a need to implement holistic programming to address uneven distributions of access to quality care [27,32]. State governments can intentionally reduce inequities by providing free SARS-CoV-2 testing sites in African-American communities [33]. In Philadelphia, organizations like the Black Doctors COVID-19 Consortium have already begun such an initiative and would benefit from additional support [34]. These stakeholders should also pressure state governments to expand Medicaid to promote wider access to care [35]. Moreover, incentivizing emission reductions in minority populations, funding green programs, and encouraging input from community members will address health concerns emerging from environmental injustice [36]. It is important to note the methodological limitations of this ecologic study. First, the data used in this report was restricted to the aforementioned timeframe and selected risk factors. Therefore, this report does not describe earlier or future descriptive trends. Second, due to the group-level nature of the data, our results are limited to description and analyses of relationships between US states and counties. While we identified a positive relationship between COVID-19 mortality and percent African-American population, the same results cannot be applied to individuals-we provide no evidence that African-Americans are more likely to die from complications resulting from COVID-19, due to the limitations of the ecologic data. Third, not all racial and ethnic groups are represented in this study, and there is no delineation between subgroups within the African-American community. Future analysis could explore the relationship between COVID-19 mortality and demographics among counties in the full US population. --- Conclusion The disproportionate effect of COVID-19 on the African-American community clearly demands action at the local, state, and US national levels. Through GIS mapping, we observed similar descriptive patterns of COVID-19 mortality and African-American population among US states and PA counties. Higher rates occurred in northeast states and eastern counties, where African-American population sizes were relatively high. These descriptive patterns were confirmed through bivariate statistical tests, and persisted after controlling for confounders. We need equity-driven public health interventions to address these disparities, including social, behavioral, economic, and policy changes prioritizing areas exhibiting the greatest burden of COVID-19 mortality, especially in minority communities. Examples include providing free SARS-CoV-2 testing in African-American communities, expanding Medicaid, and incentivizing emission reductions in minority communities. Of vital importance is the need for more comprehensive race-specific COVID-19 outcome data. With access to this data, researchers can readily identify geographic and demographic focus areas for population based prevention efforts, and improve health equity. --- Data Availability The mapping data utilized for this report were derived from the following sites: 1. U.S. Census. TIGER/Line shapefiles. 2010. https://www.census. gov/cgi-bin/geo/shapefiles/index.php?year=2010&layergroup= Counties+%28and+equivalent%29 Accessed 8 June 2020. 2. PolicyMap. Health. 2020. https://tju-policymap-com.proxy1.lib.tju. edu/maps Accessed 26 August 2020. 3. PolicyMap. Demographics. 2010. https://tju-policymap-com. proxy1.lib.tju.edu/maps Accessed 26 August 2020. --- Compliance with Ethical Standards --- Conflict of Interest The authors declare that they have no conflict of interest. Ethical Approval Not applicable. Informed Consent Not applicable. Code Availability Not applicable. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

The Centers for Disease Control and Prevention has identified African-Americans as having increased risk of COVID-19-associated mortality. Access to healthcare and related social determinants of health are at the core of this disparity. To explore the geographical links between race and COVID-19 mortality, we created descriptive maps of COVID-19 mortality rates in relation to the percentage of populations self-identifying as African-American across the USA, by state, and Pennsylvania (PA), by county. In addition, we used bivariate and logistic regression analyses to quantify the statistical relationship between these variables, and control for area-level demographic, healthcare access, and comorbidity risk factors. We found that COVID-19 mortality rates were generally higher in areas that had higher African-American populations, particularly in the northeast USA and eastern PA. These relationships were quantified through Pearson correlations showing significant positive associations at the state and county level. At the US state-level, percent African-American population was the only significant correlate of COVID-19 mortality rate. In PA at the county-level, higher percent African-American population was associated with higher COVID-19 mortality rate even after controlling for area-level confounders. More resources should be allocated to address high COVID-19 mortality rates among African-American populations.

Results The prevalence estimates of carers increased during the two decades from 3.7% in 1994 to 6.7% by 2014. Large increases in the proportion of retired carers, those aged 70 years and over, those carers employed, and those with higher educational qualifications were observed. There were also larger proportions of respondents with a country of birth other than Australia, UK, Ireland and European counties. The APC analysis illustrated an increasing prevalence rate over each decade for carers aged 20-80 years, especially for those over the age of 60 years. --- Conclusions The results illustrate changing carer characteristics and carer prevalence estimates in South Australia as new generations of carers take on the caring role. There is a need to include questions regarding informal carers within ongoing mainstream population surveys, particularly at state levels, so as to plan for their future health care and home support. --- Background Although the concept of kinship support and filial piety has existed throughout history across most cultures, the importance of family members caring for ill or aged relatives was not adequately recognised at the level of social policy of western countries until later in the twentieth century [1][2]. Traditional expectations of family meant that the caring efforts of informal carers (unpaid caregivers) were often taken for granted [3]. Today carers are more readily recognised as a separate group in their own right. Their pivotal role in health and social support systems are acknowledged for their significant economic contribution to containing health care costs [4][5]. Defining an informal carer or family caregiver is however problematic as carers can be any age, from children to the elderly, younger than nine years old to over 90 years of age. They may care for a child or adult age person with a disability, a chronic physical or mental illness, is recovering from illness or accident or who is a frail aged person. Carers are referred to as unpaid, informal or family caregivers who provide in-home support to a family member or friend who needs assistance in their daily living activities. Globally the ongoing need for an availability of informal carers is taking on even greater critical relevance. Firstly as life expectancy increases and the population ages; secondly as studies of global burden of disease show a transitioning from early mortality associated with acute fatal diseases to long term morbidity dominated by chronic conditions [6][7]. Not only have these phenomena resulted in higher proportions of disability and impairment across populations, but are impacting at community and individual levels [8]. For example informal family carers of all ages provide multifaceted care for children and adults with a disability, persons who are frail and aged, chronically ill partners and friends with complex and demanding physical and mental health problems [9]. Growing evidence suggests that the extra burden on carers can put them at risk of physical health and emotional stress and strain [10][11][12]. Therefore important as it is to track the prevalence of those conditions which influence the burden of disease, (for example cancers, cardiovascular diseases, injury as well as dementias), it is also necessary to track the prevalence, demographic profiles and health status of those who provide the informal care. Over the past thirty years, prevalence figures of informal caregiving at the population level have emerged haphazardly across industrialised nations and more recently from developing countries [13][14]. The methodology to identify informal caregivers still remains inconsistent, with international surveys providing estimates ranging from 15% to 39% [9,15]. It is acknowledged that prevalences are affected by different definitions of informal carers thus it is difficult to compare prevalences across countries and across studies or surveys. However some larger population studies do provide an overview. For example, in 2011, 12% of the British (adult) population were identified as caregivers [16]. Canadian surveys have estimated that overall 28% of adult Canadians aged 15 or over were caregivers although rates varied considerably across the provinces [17]. In the United States (US) it was estimated that up to 25-30% of the adult population were providing care and support to family and friends, but again rates varied by state [18][19]. In both the 2009 and 2012 Australian Bureau of Statistics (ABS) Disability Ageing and Carers surveys (SDAC), approximately 12% of the Australian population was identified as providing some care while approximately a third were primary carers [20][21]. ABS limit their definition to care provided to people with disability, long-term conditions, or care for persons who are aged 60 years and over [21]. At state levels in Australia, population-based details of carer prevalence have also mostly come from data collected by the national ABS surveys. These have been conducted approximately every five to six years since 1993. In the state of South Australia (SA), the prevalence of carers has been determined through additional separate state wide Health Omnibus Surveys (HOS) which have included carer status questions in nine of the annual surveys between 1994 and 2014. The aim of this paper is to show changes over 20 years, between 1994 and 2014, in the prevalence and demographic characteristics of adult carers aged 15 and over in the state of SA. To achieve this, the trend over the past decades was analysed. Secondly, multivariable analyses were conducted to determine the demographic characteristics of those reporting that they were carers, from three time points. Thirdly, the percentage differences for demographic and socio-economic variables were analysed across the two decades from 1994 to 2014. Lastly the age-period-cohort (APC) effects were also examined. The benefits of APC analysis allow the effects of age, period and cohort to be interpreted independently whilst taking into account a plethora of individual, societal, historical and cultural aspects [22]. Age relates to the physiological processes associated with growing older, period effects relate to particular time points with the assumption that populations are all equally affected, and cohort effects relate to experiences during particular time frames. --- Methods The Health Omnibus Survey is a population-based cross-sectional, representative survey that has been undertaken annually or bi-annually in SA since 1990. It investigates a range of health and health service issues as requested by health related organisations and researchers in SA and beyond. The full HOS methodology has been previously described [23] but in brief, each survey is a clustered, multi-stage, systematic, self-weighting sample selected from the Adelaide metropolitan area with the remainder being drawn from those country areas with a population of 1000 or more, based on ABS Census information. Each survey is face-to-face and interviews are undertaken by trained interviewers. --- Carer questions Informal carers are those who provide the main care in the home setting, are aged 15 years or older, and are giving ongoing personal care and assistance to dependent relatives and individuals with a chronic mental or physical illness or who are frail and aged. Providing this care is beyond that which is expected in a normal relationship [24][25]. Data pertaining to informal carer status was from selected years between 1994 to 2014. In 1994 and 1998, the carer question used was, "Are you a carer of a dependent person? (A dependent person is someone who has a chronic condition that is unlikely to improve, for example frail aged, disabled etc.). In the remaining surveys (2000-2002, 2004, 2008, 2013-2014) the carer question was, "Do you provide long-term care at home for a parent, partner, child, other relative or friend who has a disability, is frail, aged or who has a chronic mental or physical illness, where long-term care is a minimum of 6 months and may extend into years." Refer to S1 File. --- Demographic and socio-economic questions Demographic variables included in the surveys are gender, age group, area of residence, country of birth and marital status. Socio-economic variables included educational attainment and work status. Details of the annual income for each household was obtained and the socio-economic disadvantage of neighbourhood at an environmental level (using postcode) was classified into the Socio Economic Index for Areas (SEIFA) Index of Relative Socio-Economic Disadvantage from which quintiles were determined [26]. Refer to S2 File. --- Data analysis The survey data were weighted by age, gender and geographic locations so that the findings apply to the demographic profile of SA using either the ABS census data or the most recent estimated residential population for each year. Initially the trend in prevalence was determined using the nine years of data. Age and sex standardized prevalence estimates were also produced. For ease of interpretation data from three surveys collected ten years apart (1994,2004 and 2014) were selected to highlight specific demographic changes. Analysis was undertaken using SPSS Statistics, Version 19 (IBM SPSS Statistics, New York, NY, USA). For the APC analysis, the combined data from the nine years were used and an APC model was constructed using STATA Version 13 (StataCorp, College Station, TX, USA) with the 'apcfit' command [27]. APC analysis was chosen so as to interpret the effects of ageing, birth cohorts and time periods in relation to carer prevalence and odds ratios. In this analysis,'AGE' was the self-reported age of the respondents at the time of the survey interview. 'COHORT' was the age subtracted from the survey year. This ranged from the oldest respondents (80 years and over) to the youngest respondents (aged 15 years). 'PERIOD' represents the years of data collection (1994 to 2014). --- Ethical approval This procedure was approved by both the Research Ethics Committee of The University of Adelaide (U of A H-097-2010), and previously with South Australia Health and the South Australian Department of Health, (310/07/2012). Participants provided informed verbal consent for each of the surveys used between 1994 and 2014 for this study. In line with other epidemiologically-based surveillance systems, verbal consent was obtained from the participant before the interview commenced. Participation in each HOS study is voluntary. An approach letter introducing HOS was sent to selected households prior to all interviews, including a brochure outlining confidentiality and privacy assurance, how the information was to be used and which organisations were involved in each survey. If the selected respondents had any queries or did not wish to participate in the survey, they were able to call a 1800 free call telephone number listed in the introductory letter. Verbal consent was obtained by the interviewers at the time of the face-to-face interview and upon initial contact, the interviewers repeat the purpose of the survey as well as the expected length of time to complete the interview. The respondent could choose to not answer any question or section throughout the interview or could terminate participation at any time. Continued participation was taken as evidence as continued willingness to participate (implied consent) however if the participant withdrew their consent to participate at any time, the interview was terminated and no information collected. --- Results Overall, the total sample for the nine surveys was N = 26,788 (n = 1,504 were carers). The survey response rates decreased over the 20 year period from 72.4% in 1994 to 54.4% in 2014. The prevalence trend using data from nine surveys which included carer status questions is presented in Table 1 and Fig 1. In SA over the two decades from 1994 to 2014, there was almost a doubling of the prevalence of carers increasing from 3.7% 1994 to a peak of 7.9% in 2008 then declining to 6.7% by 2014. Table 2 highlights the unadjusted and adjusted odds ratios of carers over three specific years; 1994, 2004 and 2014 by a range of demographic variables. Multivariable analysis shows that in 2014 carers were more likely to be female (OR 1.43; 95% CI 1.00-2.05, p = 0.050), be aged 50-69 (OR 1.82; 95% CI 1.11-3.01 p = 0.0.019), be born in countries other than Australia, UK or Ireland, (OR 1.82; 95% CI 1.08-3.07, p = 0.02), be never married (OR 0.47; 95% CI 0.26-0.83, p = 0.011), declare their work status as 'home duties' (OR 1.88; 95% CI 1.02-3.48, p = 0.043) and have annual household incomes of $20,000-40,000 (OR 2.38; 95% CI 1.29-4.02, p = 0.005). Table 3 details the percentage increase across the two decades from 1994 to 2014 by demographic characteristics. Moderate increases across the 20 years from 1994 to 2014 were seen for females (100% increase, from 4.0% (95% CI 3.0-5.4) to 8.0% (95% CI 6.6-9.5), with a 86.2% increase in males over the same time period from 2.9% (95% CI 2.1-4.0) to 5.4% (95% CI 4.0-7.4). There was a larger 120% increase for carers aged 70 years or more, from 5.0% (95% CI 3.0-8.4) to 11.0% (95% CI 8.0-14.8). It is also noted in carers aged 15-49 years, there was a 90.9% increase from 2.2% (95% CI 1.5-3.1) to 4.2% (95% CI 3.1-5.7). Other socio-demographic increases over the 20 years included a 108.1% increase for carers with a secondary school level or less education, from 3.7% (95% CI 2.8-4.8) to 7.7% (CI 95% CI 6.2-9.6), however in terms of educational attainment, there was a 112.1% increase in carers with trade qualifications, certificates and diplomas, increasing from 3.3% (95% CI 2.2-4.9) to 7.0% (95% CI 5.3-9.0). Again, large percentage increases were recorded for work status, with a 119.0% increase for employed carers (full or part time) from 2.1% (95% CI 1.4-3.3) to 4.6% (95% CI 3.5-6.2), and a 110.9% increase in those carers nominating they were retired, from 5.5% (95% CI 3.8-7.9) to 11.6% (95% CI 9.3-14.4). Annual household income of $40,000 or more showed a 172.2% increase from 1.8% (95% CI 1.0-3.3) to 4.9% (95% CI 3.5-6.8) and a 403.3% increase for annual household income of $20-40,000 was recorded, from 3.0% (95% CI 1.9-4.8) to 15.1% (95% CI 11.5-19.6). Other demographic percentage increases included a 163.6% increase for other country of birth, increasing from 3.3% (95% CI 1.8-5.7) to 8.7% (95% CI 5.9-12.6). Carers born in Australia showed a 93.3% percentage increase from 3.0% (95% CI 2.3-3.9) to 5.8% (95% CI 4.6-7.3), however there was a smaller 43.1% increase in carers whose country of birth was UK/Ireland, from 6.5% (95% CI 4.3-9.8) to 9.3% (95% CI 6.4-13.4). In terms of percentage differences over the most recent decade, from 2004 to 2014, highest percentage increases were seen for annual household income of $20,000-40,000 (104.1% increase) from 7.4% (95% CI 5.4-10.1) to 15.1% (95% CI 11. 5-19.6). No change in the most recent decade to 2014 was found for carers in the 18-49 age group (4.2%). The only negative percentage differences occurred in the most recent decade to 2014 was for carers born in Australia (-1.7%) and carers who were never married (-14.3%). Fig 2 provides the results of the APC analysis. On the left axis is shown the independent effects of age (prevalence) and on the right axis the birth cohort and period effects, both using rate ratios are shown. The peak age for carers was around 80 years and showed a steady increasing rate over each decade for those aged 20-80 years. The graph indicates that the prevalence of being a carer increases, especially after the age of 60. In the cohort analysis the Baby Boomers born around 1951-2 are the reference point [= 1] and also represent the point of acceleration of risk which for the purposes of this paper, can be interpreted as each cohort's exposure to informal caregiving. The graph also shows a higher ratio of caring in the later cohorts born mid1970s to 2000, suggesting that persons born around 1975 (Generation X) may be twice as likely to become carers, whereas persons born 1980s to 2000 (Generation Y) have a three-fold likelihood of becoming carers. The Estimated Period Effect represents the specific calendar period when the sample population were surveyed. The resulting period effects of the graph show increasing prevalence from 1994, peaking at 2008 then falling into negative effects by 2014. --- Discussion The results indicate an inital increase in the prevalence of carers from 1994, a doubling of the proportion of carers by 2008 (3.7% to 7.9%), then a slight decline to 6.7% by 2014. Increases in proportions of carers across all ages were found, especially the 50-69 year old group and the 70 years and over age groups, a finding consistent with SA having been the state with the highest proportion of older age population in Australia [28][29][30]. In our surveys there was still a considerable proportion of carers whose education was secondary school or less, but there were higher prevalences of carers who had high educational attainment and tertiary training. Carers were more likely to be employed either full time or part time, but as would be expected with an ageing population, there were also more carers who had left work and retired. In terms of annual household incomes of carers which have increased over the past 20 years, much of this would be the result of inflation and many carers still remain in the low/ lowest most disadvantaged quintiles compared with the wider community. It is well documented that women have been the traditional and dominant family caregivers in societies of all cultures [2]. Although there was a slight increase in proportions of male carers in our earlier SA survey of 2004, overall male percentage increases (86%) remained below the 100% increase of females. ABS national surveys across Australia amongst 'all' carers in 1998 showed there was some narrowing of the gender gap with just over a half of all carers being female. In the over 75 age group there were slightly more male primary carers. Interestingly Australian disability surveys revealed that after the age of 85, carers were frequently males caring for a disabled wife [31] The implications of this changing demographic with an increasing number of male carers in this older age group raise questions about the mix and type of services needed for the future. There is a paucity of published literature on informal caregiving in terms of surveys applying Age-Period-Cohort methods. Earlier studies include one that explored changes in attitudes towards caring for ageing parents [32]. More general research examined trends in disability in older adult cohorts using APC [33]. Other APC studies and those specific to carers using methodology comparable to this paper have not been sourced to date, except those applied to prevalence rates of chronic disease topics such as diabetes, and obesity [34]. Our APC analysis (Fig 2 ) illustrates the combined impact of caregiving and ageing on several birth cohorts, which range from the older carers of pre-war generations (born early 1900s to 1945), to the Baby Boomers (born 1946-1964) and Generation X, (born 1965-1980s). The prevalence of becoming a carer showed a steady increase over each decade, especially after the age of 60. Using the APC analysis, carers' peak age was around 80 years. Again our study indicates there has been a trend for older age carers to continue caregiving into their late 70s, 80s and beyond. Other authors have observed that the onset of caregiving peaked in late middle age and older, and that informal care could span three decades or more of adult life [35]. Of concern has been the lack of data on prevalence of those much older generations who might be providing care for a spouse, relative or an adult child with a disability. Literature is more readily available on social and health impacts on those cohorts [36][37][38]. Our study highlights changes in carers' country of birth, an aspect that reflects sixty years of increased immigration to South Australia. Although our results show that there has been a 93% percent increase in carers born in Australia, by 2014 there were large percentage increases (over 163%) in carers born in countries other than Australia, UK, Ireland or Europe. This change from earlier carer profiles of the 1990s reflects the wider cultural characteristics of people (families) who have migrated to SA since the 1950s and 1960s. Australia has a rich multicultural heritage of people from over 200 countries. Those overseas-born persons aged 65 or over, have expanded in recent decades, doubling since 1991 [39][40]. There is potential for this cultural influence to impact on carers' use of home based care if there are inadequate culturally appropriate respite services or facilities. Further, there could be a reluctance on the part of older persons from different cultures to accept support services creating additional difficulties for the spouse or other family carers [41]. The results of our surveys show that new generations of carers have emerged with different characteristics in areas of education and employment and this can be a challenge for many to combine with their informal caring. The higher proportion of working carers in our study, (both full time and part time), takes on greater importance because of the many Baby Boomers who are the current generation of informal carers. This transition occurred as older Pre-World War II carers, (for example, ageing parents of the early Baby Boomers), became the recipients of care over the past 20 years. This coincided with a greater emphasis on community care in Australia during the past decades, which is being further developed as new government reforms and initiatives for carers and consumers [42]. For example they focus on consumer directed care packages based on ageing in place (and in the home) and in disability 'consumer choice and control' through the NDIS. Both models emphasise that they are consumer directed for home based care. Current integrated support services for carers are undergoing further development. Younger working carers, especially parent-carers, are another group that cannot be overlooked. Carers of younger children with chronic health conditions and disabilities are more likely to be subject to employment constraints. Employment for carers can be problematic and research is emerging of the negative effects on the parents who are the informal carers [43]. In general, of those carers who also work, many do manage well without adverse health changes, but Schofield concluded that 'working carers providing high levels of care represent a vulnerable subgroup where supportive and preventive services might be focused' [44]. In our study there was a higher proportion of working carers overall, but on a global scale, Australia has had a lower percentage of employed carers (38.4%), than other countries such as Canada, USA or the UK (55-60%) [45][46][47][48]. While the changing prevalence of carers in our study showed an increase over time, the later decline by 2014 could have been influenced by a number of factors. For example natural attrition (deaths) amongst carers from the oldest birth cohorts would have been occuring during that time. On a national scale reduced disability rates amongst both children and adults were observed which could have resulted in slightly less demand for informal carers [20]. Other factors impacting on carer numbers may be due to younger generations choosing careers over caring roles, especially with greater labour force participation of women [49]. Also it is not uncommon for people to choose to work beyond the ages of 65 and 70 years old which would lessen their availability for complex caregiving in the home. There have also been trends towards more active retirement, especially amongst baby boomers and later generations [40]. It is conjectured that some of these socio-economic factors surrounding caregiving may have been further influenced by the Global Financial Crisis (GFC) with a general reluctance to give up paid work and income during such uncertain times. Those directly affected by the GFC may have had to actively seek employment, as a priority over any caring role. The strengths of this study are that the results provide an analysis of carer prevalence and demographics over a twenty year period. It has used population-based data using face-to-face interviews, the gold standard of surveys, and a significant number of interviews were conducted. It is therefore more generalizable to other ageing populations and includes a number of relevant demographic covariates. Limitations of the study include definitional issues which continue to influence all caregiving research and make comparisons of carer / caregiver prevalence figures difficult across studies [18]. Although the data were age/sex standardized no adjustment for Consumer Price Index (CPI) was undertaken, so those household income results should be assessed with caution. There was a potential bias from survey non-response in the latter surveys in this study and this should be seen as a weakness of the project. There is a general trend towards lower response rates in all types of population surveys as people protect their privacy, or are overwhelmed by marketing telephone calls or mail outs. Due to their small numbers, non-English speaking individuals, Indigenous and those from specific cultures are included in the data collection but not in the analysis. Also the small numbers of young carers meant that meaningful analysis could not be included in our study, however this is a limitation that does not detract from their vital roles and urgent need for further research [50][51]. In conclusion, this study demonstrates there has been an overall increase in informal caregiving in SA, a state which over the past two decades has shown higher proportions of people aged 65 years and over, than those in other mainland Eastern Australian states [29]. Our research has highlighted major demographic shifts between 1994 and 2014 and it is important that policy and planning keep pace with these changes and projections. As future prevalence rates are watched with interest in SA, these findings may also be relevant in other specific populations with similar demographic profiles. If we are to sustain the current model of care in the community and the informal carers in their caring role it is important to continue monitoring the prevalence, demographic and health profiles as emerging generations of informal carers with more diverse characteristics take on the caring roles. To achieve this there is an urgent need to include informal carers within ongoing mainstream population surveys, wherever appropriate, so as to provide statistics to plan for their future health care and home support. As the literature suggests, many carers themselves have health problems or suffer diminished quality of life [52][53]. Therefore carer well-being and particularly their health status needs to be considered in parallel with those persons they are caring for. Important as it is to track the prevalence of the wide range of conditions which influence the burden of disease, it is also important to track the prevalence, demographic profiles and health status of those who provide the informal care, as many are carrying a double burden of disease. That of the care recipient, and their own. --- Author Contributions Conceptualization: AFS TKG KP RW AWT. --- Methodology: AWT RW AFS. Writing -original draft: AFS AWT. --- Writing -review & editing: AFS TKG KP RW AWT

The ongoing need for an availability of informal carers is taking on greater relevance as the global burden of disease transitions from acute fatal diseases to long term morbidity. Growing evidence suggests that extra burden on family carers may further impact on their health and ability to provide care. Important as it is to monitor the prevalence of those conditions which influence the burden of disease, it is also important to monitor the prevalence and health profiles of those who provide the informal care. The aim of this study was to demonstrate the prevalence and demographics of adult carers aged 15 and over in the state of South Australia over 20 years between 1994 and 2014.Data from nine representative, cross-sectional population surveys, conducted in South Australia, Australia were used, (total N = 26,788 and n = 1,504 carers). The adjusted prevalence estimate of carers and their demographic characteristics were determined. So as to examine whether there were any generational effects on the prevalence of carers, an Age-Period Cohort (APC) analysis was undertaken.

Introduction Non-communicable diseases (NCDs), also known as chronic diseases, result from genetic, physiological, environmental, and behavioral factors. These long-term and sometimes life-long pathologies include cardiovascular diseases, cancers, chronic respiratory diseases, diabetes as well as mental health and neurological disorders [1]. Since 2015, NCDs have surpassed infectious diseases as the leading cause of morbidity, disability, and death worldwide [2,3]; in 2019, they were accountable for 71% of all-caused deaths, the major global cause being cardiovascular diseases (17.9 million deaths), followed by cancer (9.0 million), respiratory diseases (3.8 million), and diabetes (1.6 million) [3]. Many of these premature deaths occurred before the age of 70, with most of them being preventable and treatable. While multiple interventions in diverse policy areas have been recommended to increase the prevention and control of NCDs and have achieved a noticeable reduction of their impact [1], their global economic burden remains high and increasing because of the growing age of the worldwide population [4,5]. Among all NCDs, mental illnesses and cardiovascular diseases account for the highest financial burden [6]. In particular, mental health pathologies, like major depressive disorders or anxiety, lead to a sedentary 2 of 22 lifestyle, which in turn may increase the predisposition toward other disease conditions (comorbidities). NCDs and the associated impairments and comorbidities greatly impact a patient's quality of life. The leading environmental and behavioral risk factors for NCDs are tobacco use, air pollution, physical inactivity, unhealthy diet, and the harmful use of alcohol; many individuals combine several risk factors associated with the development of NCDs [7,8]. In the management of NCDs, primary care physicians are often the first point of contact for patients [9]; relationships established between healthcare professionals (HCPs) and patients are primordial for health monitoring, providing necessary education on relevant pathologies and treatment adherence [3]. Informal carers/caregivers (a person who provides-usually-unpaid care to someone with a chronic illness, disability, or other long-lasting health or care need outside of a professional or formal framework) are also important in the management of NCDs as well as for the physical and mental well-being of the persons coping with NCDs. In December 2019, a novel coronavirus infection caused by SARS-CoV-2 first emerged in China; following its worldwide spread, the World Health Organization (WHO) declared the coronavirus disease 2019 (COVID-19) as a pandemic on 11 March 2020. Between March and May to June 2020, governments of most countries worldwide declared health emergency states and imposed confinement restrictions (lockdowns) to tackle the pandemic and avoid an overflow of intensive care units in hospitals. With the rapid spread of COVID-19 across the world, broad and sudden disruptions to health services occurred, mostly due to restrictions in public transport hindering access to health care facilities for patients, insufficient Personal Protective Equipment (PPE), insufficient clinical staff because of redeployment to COVID-19 units as well as the unavailability of essential medicines and services [10]. Those disruptions in health care might have had long-term consequences for patients living with NCDs, especially those particularly vulnerable and requiring regular or long-term care. Additionally, patients with pre-existing NCDs have been declared to be more at risk of serious illness or death from coronavirus infection [10,11]. Due to the chronic nature of their pathology, NCD patients most often need access to essential health or rehabilitation services over long periods of time. In a WHO report based on a global survey, 59% of countries reported that access to outpatient services was restricted to some degree, with rehabilitation services being the most likely impacted (50% of countries reported partial disruption and 12% complete disruption). In Europe, 79% of countries reported disruptions to rehabilitation services [10,12]. As noted by Caminiti et al., admissions for NCDs in northern Italy dropped by approximately one third in 2020 compared to 2019 [13]. Similarly, Jones et al. described a substantial decrease in screening, case identification, and referral in symptomatic cancer diagnosis in the UK for the same period, with all UK national cancer screening programs being suspended [14]. Likewise, a substantial reduction in the number of patients with acute coronary syndrome admitted to hospital in England by the end of March 2020 was observed, which likely resulted in an increase of deaths and long-term complications of myocardial infarction [15]. Additionally, a study reported that 2.3 million cancer surgeries had already been cancelled or postponed globally during the first lockdown due to COVID-19; overall, the authors predicted that 28 million elective surgeries would be cancelled or postponed worldwide in 2020 [16]. In the USA, 92% of elective vascular surgeries were cancelled at the beginning of the pandemic in April 2020 [17]. The objective of this study, performed in two independent waves, was to understand how patients with NCDs-a population that is particularly vulnerable in the context of the pandemic and lockdown-have been affected during the COVID-19 outbreak in Europe and in the USA. To do so, this study aims to evaluate: (1) the impact of COVID-19 and lockdowns on different aspects of NCD patients' lives (treatment compliance and access to medication, communication and consultation with HCPs, development of new behaviors, impact on work situation, impact on mental health), and (2) patients' expectations towards information on COVID-19 (topics of concern, most consulted and trusted sources of information, needs for additional information and services). --- Materials and Methods --- Study Design Carenity, an international online patient community, conducted a study based on an online survey among adult patients affected by NCDs. The study aimed to understand the impact of the COVID-19 pandemic on patients with NCDs in terms of the evolution of their medical condition and access to physicians and treatments, and to describe their sources of information regarding COVID-19 and their needs for specific information and support. This study was conducted in two waves: a first wave following the first strict lockdown and a second one some months later, while most countries were still under sanitary restrictions due to COVID-19 infection (strict or soft lockdown depending on the ongoing stage of the pandemic). --- Study Population The study population included registered Carenity members. Carenity is a leading digital patient platform with 400,000 patients and caregivers worldwide (88% patients/12% caregivers). Through Carenity communities, members can share experiences and find health-related information on their condition as well as contribute to medical research by participating in online studies to highlight their patient experience. Eligible patients were defined as adult patients (18 years old or older) living in Belgium, France, Germany, Italy, Spain, the UK, or the USA, having at least self-reported a diagnosed NCD and who were members of the Carenity community. The NCDs included metabolic diseases, chronic respiratory diseases, mental health diseases, cardiovascular diseases, and cancers. Because of the small sample size of the respondents in Belgium (Table 1) and their belonging to the French speaking area, Belgian patients who answered the survey on the French Carenity platform were reported together with the French population for the purposes of the study. * Percentages may not equal 100 because of rounding; a >1,000,000 inhabitants; b 100,000 to 1,000,000 inhabitants; c 20,000 to 100,000 inhabitants; d 2000 to 20,000 inhabitants; e <unk>2000 inhabitants; f Including children. Patients meeting the inclusion criteria completed an initial survey (wave 1) and/or a second survey (wave 2) to self-report the impact of the COVID-19 pandemic and lockdown on their care pathway. The two survey periods were enrolled independently, and patients who participated in wave 1 could self-select, but were not required, to complete the wave 2 survey. The study population was defined by all patients who completed the questionnaire (wave 1 and/or wave 2). --- Ethical Considerations This study was conducted in accordance with the Declaration of Helsinki and following the principles of good clinical practice. Prior to data collection, all patients gave their informed online consent (patients were informed that the treatment of their health data is conducted based on their explicit consent, formalized by the click of the "Start" button at the bottom of the survey's information page, and clicking on the button), which allowed for data collection and publication. The participants' privacy and confidentiality was guaranteed following the USA health care personal data protection law (HIPAA, Health Insurance Portability and Accountability Act) as well as European laws and regulations (GDPR, General Data Protection Regulation). Ethical review and approval were waived for this study, as the aim of the study is to determine the patients' insights from their experiences, and informed consent was given by all patients prior to answering the survey. --- Data Collection Carenity has setup a barometer (the 2 independent waves) to measure the impact of COVID-19 on access to health care and quality of life for chronic patients in real time. Members of the Carenity NCDs communities were asked to complete the same online survey at two distinct periods. In the countries involved in this study, a strict lockdown was in place for several weeks between March and June 2021, depending on the country; from 22 June 2020 to 14 August, patients reported experience during this first lockdown in Europe and in the USA (data from wave 1). The data regarding the impact of the second lockdown were collected in real time (prospectively) from 2 November 2020 to 29 December 2020 (data from wave 2). During the second survey sample (wave 2), not all of the countries included in this study had a further strict lockdown at that point in time; however, all of the maintained strict sanitary restrictions concerning displacements and physical and social distancing were still in practice in those countries. During the second wave, many EU countries were considering a stricter lockdown, which began to be applied during the yearend holidays. The collected data of both waves were analyzed in two separated cohorts; patients could participate in both waves, but this was not mandatory. For comparison purposes, the same survey including 39 closed-ended questions was used for both waves. The questionnaire was validated to ensure compliance for each country involved as well as by native speakers, as it was programmed in local languages for the different Carenity worldwide platforms. The collected data included self-reported sociodemographic and medical profiles as well as information about the patients' health-related quality of life (HR-QoL) during the COVID-19 pandemic and lockdown and their needs towards the pandemic. The following notions were especially evaluated: disability (disability is characterized by the limitations or reductions in the individual's ability to participate in their activities of daily living due to a disease or health condition) [18], social isolation (social isolation is the inadequate quality and quantity of social relations with other people at the different levels where human interaction takes place (individual, group, community and the larger social environment)) [19], depressive symptoms (depressive symptoms can vary from mild to severe and can include: (1) feeling sad or having a depressed mood, (2) loss of interest or pleasure in activities once enjoyed, (3) changes in appetite-weight loss or gain unrelated to dieting, (4) trouble sleeping or sleeping too much, (5) loss of energy or increased fatigue, (6) increase in purposeless physical activity (e.g., the inability to sit still, pacing, handwringing) or slowed movements or speech (these actions must be severe enough to be observable by others), (7) feeling worthless or guilty, (8) difficulty thinking or concentrating or making decisions, and (9) thoughts of death or suicide [20]), and regular treatment/long term treatment (there is no commonly accepted definition of long-term treatment. In our survey, it has been defined as a treatment that was taken for at least three months. No differences have been made between regular and long-term treatment). Data collected on the Carenity platform are hosted in France on a secured computer server in accordance with the requirements of the "Commission Nationale de l'Informatique et des Libertés" (CNIL)-declaration number n • 1484083, dated 29 March 2011. --- Statistical Analysis Descriptive univariate and multivariate analyses were performed using the statistical software R (version 3.5.2 (The R Foundation, Vienna, Austria)). For quantitative data, the following indicators were calculated: mean, range (minimum, maximum), frequency-n (%). Multivariate analysis included subgroup analyses by country. A multiple correspondence analysis was also performed using the statistical software R (version 3.5.2) for the clustering analysis, and internal validation measures were used to assess the quality of the clustering. The resulting clusters were thus analyzed through a hierarchical cluster analysis. The p-values were calculated with a confidence interval of 95%. A p value <unk> 0.05 was considered statistically significant. --- Results --- Characteristics of the Study Population Overall, 22,372 members of the Carenity NCDs communities were invited to participate in the study after the first lockdown (wave 1) from 22 June 2020 to 14 August 2020 and 23,412 after the second lockdown (wave 2) from 2 November 2020 to 29 December 2020 (Figure 1). Among the 4352 patients who started the questionnaire in the first wave (3774 in wave 2), the respondent population consisted of 2489 patients (2342 in wave 2) who filled it completely with valid answers; 195 patients answered both waves.. Environ. Res. Public Health 2021, 18, x FOR PEER REVIEW 6 of 22 2020 (Figure 1). Among the 4352 patients who started the questionnaire in the first wave (3774 in wave 2), the respondent population consisted of 2489 patients (2342 in wave 2) who filled it completely with valid answers; 195 patients answered both waves. As few differences were observed between results from wave 1 and wave 2, the results of wave 1 are presented in the article, and only the differences found between both waves are highlighted for clarity. --- Sociodemographic Profile In wave 1, a total of 2489 patients with NCDs living in six different countries answered the survey for the Carenity community. Table 1 summarizes the sociodemographic profile of patients. Most patients were women (n = 1602/2489, 64%) and the mean age was 53 years (range, 18-91), with 66% (n = 1649/2489) of the patients being under 60 years old. During the lockdown, 80% (n = 1957/2489) of respondents were living in Europe (25% (n = 615/2489) in France or Belgium, 18% (n = 458/2489) in the UK, 14% (n = 354/2489) in Spain, 13% (n = 313/2489) in Italy, 10% (n = 247/2489) in Germany, and 20% (n = 502/2489) in the USA. Overall, respondents were predominantly living in small cities (n = 637/2489, 26%), followed by mid-sized cities (n = 568/2489, 23%) or metropolitan cities (n = 542/2489, 22%). During the pandemic, most participants (n = 1418/2489, 56%) were residing in a house with a yard while 33% (n = 814/2489) lived in an apartment, 5% (n = 125/2489) in a house without a yard, and 3% (n = 63/2489) in a studio apartment. Including children, most households housed two persons (n = 944/2489, 38%) during the pandemic and 24% (n = 598/2489) of households included more than three persons. Sociodemographic profiles from both surveys showed a very good degree of consistency between the NCD population in both waves. --- Clinical Characteristics Table 2 presents the clinical characteristics of the respondents in both waves. When asked about their chronic illness, 42% of participants of wave 1 (n = 1042/2489) answered As few differences were observed between results from wave 1 and wave 2, the results of wave 1 are presented in the article, and only the differences found between both waves are highlighted for clarity. --- Sociodemographic Profile In wave 1, a total of 2489 patients with NCDs living in six different countries answered the survey for the Carenity community. Table 1 summarizes the sociodemographic profile of patients. Most patients were women (n = 1602/2489, 64%) and the mean age was 53 years (range, 18-91), with 66% (n = 1649/2489) of the patients being under 60 years old. During the lockdown, 80% (n = 1957/2489) of respondents were living in Europe (25% (n = 615/2489) in France or Belgium, 18% (n = 458/2489) in the UK, 14% (n = 354/2489) in Spain, 13% (n = 313/2489) in Italy, 10% (n = 247/2489) in Germany, and 20% (n = 502/2489) in the USA. Overall, respondents were predominantly living in small cities (n = 637/2489, 26%), followed by mid-sized cities (n = 568/2489, 23%) or metropolitan cities (n = 542/2489, 22%). During the pandemic, most participants (n = 1418/2489, 56%) were residing in a house with a yard while 33% (n = 814/2489) lived in an apartment, 5% (n = 125/2489) in a house without a yard, and 3% (n = 63/2489) in a studio apartment. Including children, most households housed two persons (n = 944/2489, 38%) during the pandemic and 24% (n = 598/2489) of households included more than three persons. Sociodemographic profiles from both surveys showed a very good degree of consistency between the NCD population in both waves. --- Clinical Characteristics Table 2 presents the clinical characteristics of the respondents in both waves. When asked about their chronic illness, 42% of participants of wave 1 (n = 1042/2489) answered metabolic diseases, 37% (n = 930/2489) respiratory diseases, 24% (n = 603/2489) cardiovascular diseases, and 23% (n = 577/2489) mental disorders. Among the most frequent conditions, patients mentioned type 2 diabetes (n = 832/2489, 33%), arterial hypertension (n = 575/2489, 23%), asthma (n = 562/2489, 23%), COPD (n = 459/2489, 18%), or depression (n = 412/2489, 17%). Compared to wave 1, a decrease in the number of cited respiratory diseases (n = 772/2342, 33% in wave 2 vs. n = 930/2489, 37% in wave 1) and an increased in the number of respondents with mental disorders (n = 597/2342, 25% in wave 2 vs. n = 577/2489, 23% in wave 1) was observed during wave 2. The noted increase in the proportion of mental disorders between the first and second lockdowns (from 23% to 30% of respondents) showed great disparity among the countries: an increase of 21% took place between the first and second lockdowns in the UK (n = 42/548, 9% vs. n = 104/346, 30% of mental disorders), 11% in France/Belgium (n = 56/615, 9% vs. n = 113/565, 20%), and 9% in Germany (n = 63/247, 26% vs. 127/363, 35%), whereas mostly no change was observed in the USA (n = 200/502, 40% vs. 157/412, 38%) or in Spain (n = 109/354, 31% vs. n = 108/374, 29%). Overall, 45% (n = 1132/2489) of respondents considered that their chronic disease was disabling (disabling level of 6-10, 10 being the highest score = "very disabling"), with a mean disability level of 5.5 (range, 0-10). The second wave did not show any relevant difference compared to the first wave concerning this parameter. In total, 90% (n = 2242/2489) of the respondents were currently taking regular/longterm treatments for their primary chronic illness. Before the first lockdown caused by the COVID-19 pandemic, 41% (n = 1033/2489) of participants declared having experienced depression symptoms and 46% (n = 1138/2489) experienced anxiety. The proportion of respondents with depression symptoms increased slightly before the second lockdown (n = 1072/2342, 46% vs. n = 1033/2489, 41%); the proportion of patients with anxiety was stable (n = 1105/2342, 47% vs. n = 1138/2489, 46%). During the first lockdown, 25% of respondents (n = 612/2489) respondents had a COVID-19 test. Of those tested, 4% (n = 109/2489) had a positive result. Not surprisingly, those numbers increased during the second lockdown: among the 35% (n = 825/2342) of tested patients, 8% (n = 176/2342) revealed to be positive for COVID-19. During the first lockdown, while the most reported chronic conditions concerned respiratory diseases (n = 277/458, 60%) in the UK, metabolic diseases were the most cited answer in France/Belgium (n = 335/615, 54%), in the USA (n = 232/502, 46%), in Germany (n = 110/247, 45%), and in Spain (n = 149/354, 42%) (Figure 2). No 247 (10) 228 ( During the first lockdown, while the most reported chronic conditions concerned respiratory diseases (n = 277/458, 60%) in the UK, metabolic diseases were the most cited answer in France/Belgium (n = 335/615, 54%), in the USA (n = 232/502, 46%), in Germany (n = 110/247, 45%), and in Spain (n = 149/354, 42%) (Figure 2). --- Impact of COVID-19 Pandemic on the Medical Condition of Patients with NCDs One out of two respondents (n = 1242/2489, 50%) experienced a worsening of their medical condition during the lockdowns, the main reasons reported being a reduction or stoppage of physical activities (n = 463/2489, 19%) followed by the normal evolution of their chronic illness (n = 317/2489, 13%) and the fact that they did not consult their doctor (n = 203/2489, 8%). Moreover, 17% of respondents (n = 412/2489) developed a new disease during the first lockdown. Among those participants, 83% (n = 340/412) contacted an HCP about it: their general practitioner (for 51% of patients (n = 210/412)), their specialist (pneumologist, diabetologist, etc.) (n = 111/412, 27%), or a healthcare professional in a hospital --- Impact of COVID-19 Pandemic on the Medical Condition of Patients with NCDs One out of two respondents (n = 1242/2489, 50%) experienced a worsening of their medical condition during the lockdowns, the main reasons reported being a reduction or stoppage of physical activities (n = 463/2489, 19%) followed by the normal evolution of their chronic illness (n = 317/2489, 13%) and the fact that they did not consult their doctor (n = 203/2489, 8%). Moreover, 17% of respondents (n = 412/2489) developed a new disease during the first lockdown. Among those participants, 83% (n = 340/412) contacted an HCP about it: their general practitioner (for 51% of patients (n = 210/412)), their specialist (pneumologist, diabetologist, etc.) (n = 111/412, 27%), or a healthcare professional in a hospital emergency center (n = 75/412, 18%). Compared to the first lockdown, more specialists (151/381, 40% for wave 2 vs. n = 111/412, 27% for wave 1) and pharmacists (n = 47/381, 12% for wave 2 vs. n = 29/412, 7% for wave 1) were contacted by the respondents because of a new condition that had developed during the second lockdown. --- Impact of COVID-19 Pandemic on the Behaviors of Patients with NCDs In total, 92% (n = 2296/2489) of participants implemented solutions to take care of their medical condition during the first lockdown. Most of them (2057/2489, 83%) followed recommendations from health authorities, while 31% (n = 764/2489) of respondents scheduled a teleconsultation to avoid having to go to the hospital or the doctor's office, and 16% (n = 408/2489) used an app on their smartphone, an alarm, or a pill organizer to remind them to take their medications. The percentage of patients using teleconsultation decreased to 23% (n = 536/2342) during the second wave. Additionally, 10% (n = 240/2489) of participants mentioned that an informal (non-professional) carer had helped them. Between the first and the second lockdowns, a reduction in the compliance of the health authority recommendations was observed (n = 2, 057/2489, 83% for wave 1 vs. n = 1611/2342, 69% for wave 2). As shown in Figure 3a, 64% (n = 1601/2489) of respondents adopted a negative behavior during the lockdowns: 39% (n = 967/2489) of them reduced or stopped physical activities, 22% increased or started unhealthy nutrition (n = 539/2489), 16% increased or started taking medicine to sleep at night because of insomnia (n = 403/2489), 15% increased or started taking medicines for anxiety or depression (n = 367/2489) while some patients started or increased smoking (n = 352/2489, 14%) or alcohol (306/2489, 12%). On the contrary, half of respondents (n = 1231/2489) adopted at least one positive behavior during the lockdowns: 21% (n = 513/2489) practiced breathing exercises, 20% (n = 492/2489) increased or started healthy nutrition, 17% (n = 424/2489) increased or started physical activity inside or outside, and 12% (n = 299/2489) practiced yoga or meditation (Figure 3b). In total, 36% of patients (n = 891/2480) adopted simultaneously negative and positive behaviors, for instance, the 5% (n = 47/891) of participants who increased/started healthy nutrition and reduced/stopped sports and physical activities during the same period. emergency center (n = 75/412, 18%). Compared to the first lockdown, more specialists (151/381, 40% for wave 2 vs. n = 111/412, 27% for wave 1) and pharmacists (n = 47/381, 12% for wave 2 vs. n = 29/412, 7% for wave 1) were contacted by the respondents because of a new condition that had developed during the second lockdown. --- Impact of COVID-19 Pandemic on the Behaviors of Patients with NCDs In total, 92% (n = 2296/2489) of participants implemented solutions to take care of their medical condition during the first lockdown. Most of them (2057/2489, 83%) followed recommendations from health authorities, while 31% (n = 764/2489) of respondents scheduled a teleconsultation to avoid having to go to the hospital or the doctor's office, and 16% (n = 408/2489) used an app on their smartphone, an alarm, or a pill organizer to remind them to take their medications. The percentage of patients using teleconsultation decreased to 23% (n = 536/2342) during the second wave. Additionally, 10% (n = 240/2489) of participants mentioned that an informal (non-professional) carer had helped them. Between the first and the second lockdowns, a reduction in the compliance of the health authority recommendations was observed (n = 2, 057/2489, 83% for wave 1 vs. n = 1611/2342, 69% for wave 2). As shown in Figure 3a, 64% (n = 1601/2489) of respondents adopted a negative behavior during the lockdowns: 39% (n = 967/2489) of them reduced or stopped physical activities, 22% increased or started unhealthy nutrition (n = 539/2489), 16% increased or started taking medicine to sleep at night because of insomnia (n = 403/2489), 15% increased or started taking medicines for anxiety or depression (n = 367/2489) while some patients started or increased smoking (n = 352/2489, 14%) or alcohol (306/2489, 12%). On the contrary, half of respondents (n = 1231/2489) adopted at least one positive behavior during the lockdowns: 21% (n = 513/2489) practiced breathing exercises, 20% (n = 492/2489) increased or started healthy nutrition, 17% (n = 424/2489) increased or started physical activity inside or outside, and 12% (n = 299/2489) practiced yoga or meditation (Figure 3b). In total, 36% of patients (n = 891/2480) adopted simultaneously negative and positive behaviors, for instance, the 5% (n = 47/891) of participants who increased/started healthy nutrition and reduced/stopped sports and physical activities during the same period. --- Impact of COVID-19 Pandemic on Mental Health of Patients with NCDs During the first lockdown, 54% (n = 1330/2489) of respondents mentioned that they felt very or completely socially isolated during the lockdown (answered levels between 6 and 10 on an increasing 10-step scale, with 10 = "completely isolated") (Figure 4a); the mean level of social isolation reached 6.2 out of 10.0 (range, 0-10). In the UK and USA, respondents reported higher social isolation levels than in Europe (mean of 6.9 and 6.7/10.0 vs. 6.0/10.0, respectively) (p-value <unk> 0.05). When asked about their stress level and state of mind, 54% (n = 1346/2489) of respondents mentioned that both parameters had been strongly impacted by the lockdown (a rate higher or equal to 7 on a 10-step scale, with 10 = "extremely stressed"), with a mean stress level of 6.3 out of 10.0 (range: 0-10) (Figure 4b). A trend toward a higher impact of the pandemic on stress level and state of mind was observed in the USA (mean, 6.7) and in the UK (mean, 6.5) compared to Europe (mean, 6.2), although in Spain, a mean stress level and state of mind of 6.9 out of 10.0 was reported (p-value <unk> 0.05). --- Impact of COVID-19 Pandemic on Mental Health of Patients with NCDs During the first lockdown, 54% (n = 1330/2489) of respondents mentioned that they felt very or completely socially isolated during the lockdown (answered levels between 6 and 10 on an increasing 10-step scale, with 10 = "completely isolated") (Figure 4a); the mean level of social isolation reached 6.2 out of 10.0 (range, 0-10). In the UK and USA, respondents reported higher social isolation levels than in Europe (mean of 6.9 and 6.7/10.0 vs. 6.0/10.0, respectively) (p-value <unk> 0.05). When asked about their stress level and state of mind, 54% (n = 1346/2489) of respondents mentioned that both parameters had been strongly impacted by the lockdown (a rate higher or equal to 7 on a 10-step scale, with 10 = "extremely stressed"), with a mean stress level of 6.3 out of 10.0 (range: 0-10) (Figure 4b). A trend toward a higher impact of the pandemic on stress level and state of mind was observed in the USA (mean, 6.7) and in the UK (mean, 6.5) compared to Europe (mean, 6.2), although in Spain, a mean stress level and state of mind of 6.9 out of 10.0 was reported (p-value <unk> 0.05). During the first lockdown, the feelings of social isolation of the respondents tended to be higher in the USA (mean, 6.7 out of 10.0) compared to European countries (mean, 6.0 out of 10.0); similar outcomes have been observed in the impact of the pandemic on the stress level and the state of mind of patients (p-value <unk> 0.05) (Figure 5). This difference increased between the first and second lockdowns (Figure 5). During the first lockdown, the feelings of social isolation of the respondents tended to be higher in the USA (mean, 6.7 out of 10.0) compared to European countries (mean, 6.0 out of 10.0); similar outcomes have been observed in the impact of the pandemic on the stress level and the state of mind of patients (p-value <unk> 0.05) (Figure 5). This difference increased between the first and second lockdowns (Figure 5). Among the patients who experienced depressive symptoms before the pandemic, these symptoms worsened during the lockdowns (mean, 6.2/10.0 with 10 = "Extremely worsened"; range: 0-10). Similarly, in respondents who previously had anxiety, this symptom increased during lockdown (mean, 6.5/10.0; range: 0-10). Younger patients (<unk>40 years) tended to experience a worsening of their depressive symptoms and anxiety compared to the older ones (>50 years) (p-value <unk> 0.05). This result was observed in both waves. In addition to depression and anxiety symptoms, 19% of respondents (n = 475/2489) reported having developed a mental health problem during the first lockdown and 23% (n = 536/2489) during the second lockdown. Patients from the USA tended to be much more affected than patients living in France/Belgium (n = 127/502, 25% vs. n = 70/615, 11% in the wave 1 and n = 134/412, 33% vs. n = 82/565, 15% in wave 2). Among the patients who experienced depressive symptoms before the pandemic, these symptoms worsened during the lockdowns (mean, 6.2/10.0 with 10 = "Extremely worsened"; range: 0-10). Similarly, in respondents who previously had anxiety, this symptom increased during lockdown (mean, 6.5/10.0; range: 0-10). Younger patients (<unk>40 years) tended to experience a worsening of their depressive symptoms and anxiety compared to the older ones (>50 years) (p-value <unk> 0.05). This result was observed in both waves. In addition to depression and anxiety symptoms, 19% of respondents (n = 475/2489) reported having developed a mental health problem during the first lockdown and 23% (n = 536/2489) during the second lockdown. Patients from the USA tended to be much more affected than patients living in France/Belgium (n = 127/502, 25% vs. n = 70/615, 11% in the wave 1 and n = 134/412, 33% vs. n = 82/565, 15% in wave 2). --- Impact of COVID-19 Pandemic on the Care Pathway of Patients with NCDs During the first lockdown, patients with NCDs had to change the frequency of their visits to their physician: 38% (n = 958/2489) of respondents said they were seeing their physician less than usual, whereas 16% (n = 393/2489) reported having consulted them more than usual. Compared to the first lockdown, the proportion of patients who consulted their physician less than usual decreased from 38% (n = 958

An international online patient community, Carenity, conducted a patient study in two independent waves among adults affected by non-communicable diseases (NCDs) in Europe and in the United States of America (USA). The study aimed to assess the real time impact of the coronavirus disease 2019 (COVID-19) on the medical conditions of patients with NCDs, their access to health care, and their adaptation to daily life as well as to describe their sources of information on COVID-19 and their needs for specific information and support. During the pandemic, 50% of the patients reported a worsening of their medical condition, and 17% developed a new disease. Additionally, 26% of the respondents reported an impact of the pandemic on regular/long-term treatment intake. 54% of the patients felt very or completely socially isolated and reported a strong impact of the COVID-19 pandemic on their stress level and state of mind, with higher levels observed in the USA compared to Europe. 59% of the respondents wished to have received additional information regarding the risks associated to their medical condition during the pandemic. Television was the most used source of information, whereas physicians were the most trusted one. This study describes the substantial impact of the COVID-19 pandemic on NCD patients.

<unk> 0.05). This result was observed in both waves. In addition to depression and anxiety symptoms, 19% of respondents (n = 475/2489) reported having developed a mental health problem during the first lockdown and 23% (n = 536/2489) during the second lockdown. Patients from the USA tended to be much more affected than patients living in France/Belgium (n = 127/502, 25% vs. n = 70/615, 11% in the wave 1 and n = 134/412, 33% vs. n = 82/565, 15% in wave 2). --- Impact of COVID-19 Pandemic on the Care Pathway of Patients with NCDs During the first lockdown, patients with NCDs had to change the frequency of their visits to their physician: 38% (n = 958/2489) of respondents said they were seeing their physician less than usual, whereas 16% (n = 393/2489) reported having consulted them more than usual. Compared to the first lockdown, the proportion of patients who consulted their physician less than usual decreased from 38% (n = 958/2489) to 23% (n = 546/2342) during the second lockdown. Overall, almost 30% of respondents (n = 749/2489) declared having had difficulties in finding an available physician during the first lockdown. Moreover, medical visits or procedures were strongly impacted, as 52% of respondents (n = 1298/2489) reported that longplanned medical consultations or surgeries had been canceled or rescheduled since the start of the pandemic. However, between both waves of the study, the proportion of cancelled or postponed medical consultations or procedures decreased from 52% (n = 1298/2489) to 32% (n = 740/2342). During the first lockdown, 26% (n = 572/2242) of respondents under a regular/longterm treatment reported an impact of the pandemic on their regular/long-term treatment --- Impact of COVID-19 Pandemic on the Care Pathway of Patients with NCDs During the first lockdown, patients with NCDs had to change the frequency of their visits to their physician: 38% (n = 958/2489) of respondents said they were seeing their physician less than usual, whereas 16% (n = 393/2489) reported having consulted them more than usual. Compared to the first lockdown, the proportion of patients who consulted their physician less than usual decreased from 38% (n = 958/2489) to 23% (n = 546/2342) during the second lockdown. Overall, almost 30% of respondents (n = 749/2489) declared having had difficulties in finding an available physician during the first lockdown. Moreover, medical visits or procedures were strongly impacted, as 52% of respondents (n = 1298/2489) reported that long-planned medical consultations or surgeries had been canceled or rescheduled since the start of the pandemic. However, between both waves of the study, the proportion of cancelled or postponed medical consultations or procedures decreased from 52% (n = 1298/2489) to 32% (n = 740/2342). During the first lockdown, 26% (n = 572/2242) of respondents under a regular/longterm treatment reported an impact of the pandemic on their regular/long-term treatment intake: 6% (n = 136/2242) completely stopped some of their treatments and 10% temporarily stopped some of them (n = 230/2242), while 11% (n = 242/2242) took them more regularly than they usually did (Figure 6). Additionally, 18% of respondents (n = 407/2242) reported having difficulties in finding their treatment at the pharmacy during the lockdown. intake: 6% (n = 136/2242) completely stopped some of their treatments and 10% temporarily stopped some of them (n = 230/2242), while 11% (n = 242/2242) took them more regularly than they usually did (Figure 6). Additionally, 18% of respondents (n = 407/2242) reported having difficulties in finding their treatment at the pharmacy during the lockdown. --- Access and Need for Disease-Specific Information In all, 74% of patients had concerns about the pandemic, with 21% (n = 503/2489) of them feeling frustrated most of the time. Considering the access to specific information related to their chronic illness in the context of the pandemic, 31% of patients (n = 776/2489) did not receive any illness information, and 25% of them (n = 614/2489) reported receiving insufficient information. Of those participants, 59% (n = 822/1390) wished they had received additional information regarding the risks associated to their medical condition, while 32% (n = 445/1390) requested advice on how to deal with the end of the lockdown, and 29% (n = 408/1390) needed more information on available psychological support. During both waves, patients living in the USA tended to be more satisfied regarding the access to specific information related to their primary chronic illness compared to patients from Europe (Figure 7). --- Access and Need for Disease-Specific Information In all, 74% of patients had concerns about the pandemic, with 21% (n = 503/2489) of them feeling frustrated most of the time. Considering the access to specific information related to their chronic illness in the context of the pandemic, 31% of patients (n = 776/2489) did not receive any illness information, and 25% of them (n = 614/2489) reported receiving insufficient information. Of those participants, 59% (n = 822/1390) wished they had received additional information regarding the risks associated to their medical condition, while 32% (n = 445/1390) requested advice on how to deal with the end of the lockdown, and 29% (n = 408/1390) needed more information on available psychological support. During both waves, patients living in the USA tended to be more satisfied regarding the access to specific information related to their primary chronic illness compared to patients from Europe (Figure 7). intake: 6% (n = 136/2242) completely stopped some of their treatments and 10% temporarily stopped some of them (n = 230/2242), while 11% (n = 242/2242) took them more regularly than they usually did (Figure 6). Additionally, 18% of respondents (n = 407/2242) reported having difficulties in finding their treatment at the pharmacy during the lockdown. --- Access and Need for Disease-Specific Information In all, 74% of patients had concerns about the pandemic, with 21% (n = 503/2489) of them feeling frustrated most of the time. Considering the access to specific information related to their chronic illness in the context of the pandemic, 31% of patients (n = 776/2489) did not receive any illness information, and 25% of them (n = 614/2489) reported receiving insufficient information. Of those participants, 59% (n = 822/1390) wished they had received additional information regarding the risks associated to their medical condition, while 32% (n = 445/1390) requested advice on how to deal with the end of the lockdown, and 29% (n = 408/1390) needed more information on available psychological support. During both waves, patients living in the USA tended to be more satisfied regarding the access to specific information related to their primary chronic illness compared to patients from Europe (Figure 7). Beyond the need for more reliable and complete information from health authorities (n = 1192/2489, 48% of respondents) or in the media (n = 794/2489, 32%), respondents also expressed their needs in terms of solutions and services to better cope with the COVID-19 pandemic. Patients wished to have had unrestricted access to their family or close friends during the lockdown (n = 660/2489, 27%), access to monthly consultations with their specialists (n = 587/2489, 24%), or easier access to telemedicine (n = 579/2489, 23%). As shown on Figure 8, though television was used by 61% of patients (n = 1512/2489) as a source of information, it was poorly trusted (only by 21% of respondents, n = 532/2489). On the contrary, physicians were the most trusted source of information (n = 1153/2489, 46%) but could be used more often as they were as the fifth most used source by the patients. This outcome did not significantly change between both lockdown situations. Beyond the need for more reliable and complete information from health authorities (n = 1192/2489, 48% of respondents) or in the media (n = 794/2489, 32%), respondents also expressed their needs in terms of solutions and services to better cope with the COVID-19 pandemic. Patients wished to have had unrestricted access to their family or close friends during the lockdown (n = 660/2489, 27%), access to monthly consultations with their specialists (n = 587/2489, 24%), or easier access to telemedicine (n = 579/2489, 23%). --- Impact of the Pandemic and Patients' Need Depending on Patients' Profile As shown on Figure 8, though television was used by 61% of patients (n = 1512/2489) as a source of information, it was poorly trusted (only by 21% of respondents, n = 532/2489). On the contrary, physicians were the most trusted source of information (n = 1153/2489, 46%) but could be used more often as they were as the fifth most used source by the patients. This outcome did not significantly change between both lockdown situations. Beyond the need for more reliable and complete information from health authorities (n = 1192/2489, 48% of respondents) or in the media (n = 794/2489, 32%), respondents also expressed their needs in terms of solutions and services to better cope with the COVID-19 pandemic. Patients wished to have had unrestricted access to their family or close friends during the lockdown (n = 660/2489, 27%), access to monthly consultations with their specialists (n = 587/2489, 24%), or easier access to telemedicine (n = 579/2489, 23%). As shown on Figure 8, though television was used by 61% of patients (n = 1512/2489) as a source of information, it was poorly trusted (only by 21% of respondents, n = 532/2489). On the contrary, physicians were the most trusted source of information (n = 1153/2489, 46%) but could be used more often as they were as the fifth most used source by the patients. This outcome did not significantly change between both lockdown situations. --- Impact of the Pandemic and Patients' Need Depending on Patients' Profile --- Impact of the Pandemic and Patients' Need Depending on Patients' Profile To identify the groups of patients with similar patterns of responses to the questions, a clustering analysis on the categorical data was performed through a multivariate analysis. Based on the answers collected through the questionnaire, three different clusters emerged for each wave: younger highly impacted patients, middle-aged moderately impacted patients, and less impacted older patients (Table 3) (p-value <unk> 0.05). In comparison to Cluster 3 (older patients), respondents in Clusters 1 and 2 suffered more depressive and anxiety symptoms before lockdown and had self-reported disabling medical conditions, felt isolated and stressed, had a worsening of their medical condition, had seen their physician more than usual, were very affected by the cancellation of their consultations or surgeries, and developed a new disease (including a new mental illness) more frequently) (p-value <unk> 0.05). Moreover, during the first lockdown, respondents in Cluster 2 were those who were the least satisfied with the access to specific information related to their chronic pathology (63% unsatisfaction vs. 49% in Cluster 1 and 55% in Cluster 3). During the second lockdown, no major differences were observed in these three clusters, except that patients from Cluster 2 reported mental diseases as primary chronic illness instead of respiratory diseases and cancers. --- Discussion This online patient survey conducted via the Carenity online communities among adults self-reporting as having an NCD showed how the disruptions in health services due to the pandemic strongly affected the daily life of this exposed population. Statistical methods involving clustering analysis, social desirability, and influence given the general increased awareness of people with existing NCDs being more at risk during the pandemic. With the rapid spread of COVID-19 across the world, patients followed health authority recommendations and adapted their behaviors to the new situation (lockdown). A majority of the patients adopted negative behaviors such as reducing or stopping physical activities, starting or increasing unhealthy nutrition, and increasing smoking or alcohol use. Similar negative behaviors have been already reported in numerous European countries [21][22][23][24][25]. Compared to an active lifestyle, sedentary living and its associated comorbidities-for instance, obesity-have been linked to an increased risk of NCD development [3]. Social isolation was mostly enforced through governmental policies on social distancing; in our study, social isolation affected a lot of patients (54%) living with NCDs, especially in the UK and USA, and many patients wished to have unrestricted access to their family and close friends during lockdown. It should be noted, though that if social isolation had been largely overlooked in general, specific initiatives were developed to address this issue, for instance, the UK Campaign to End Loneliness [26]. Nevertheless, social isolation might have also been self-imposed by NCD patients because of the fear of contagion in healthcare settings-this population was described as more vulnerable to COVID-19 infection in comparison to the overall population [27]-as well as an initial period of uncertainty without an understanding of COVID-19 infection [10,11]. Increased feelings of loneliness were described, particularly in older individuals who were at a higher risk for developing mental health disorders [28]. Social isolation might have additionally increased the stress status of patients living with NCDs, as 54% of respondents reported an increased stress level during the lockdowns. Consequently, during the lockdowns, depressive symptoms and anxiety increased 51% and 56%, respectively, in respondents living with NCDs, especially in the younger population. In comparison, in a large literature review including 16 studies and covering the first seven months of the pandemic, Lakhan et al. described the prevalence of depression (20%), anxiety (35%), and stress level (53%) in the general population [29]. A Brazilian study described that patients with depression were at a higher risk for the incidence of unhealthy diet behavior during lockdown [30]. In the pandemic context, pre-existing mental disorders worsened among respondents, and additional new mental disorders appeared. Similar outcomes in terms of increased NCD risk factors and worsened clinical symptoms have been also described by Palmer et al. [28]. Our study raises the question of the long-term consequences of medical conditions worsening, such as the suicide arising in this endangered population. Additionally, the COVID-19 pandemic had a strong impact on patients' access to HCPs and their treatments. Most medical practices as well as health care and rehabilitation services suddenly closed or scaled down non-urgent visits while hospitals were overwhelmed with severe COVID-19 cases, lacking staff, PPE, and essential medicines. Additionally, confinement restrictions strongly limited the access of NCD patients to health care and medication, which increased their uncertainty. Especially in the early days of the lockdown, this situation, which led NCD patients unattended, forced them to change their access to care. In our study, respondents reported seeing their physician less than usual, having difficulties finding an available physician, and trouble finding their treatment at the pharmacy during the first lockdown. Subsequently, patients often modified the intake frequency of their regular or long-term treatment, which might have decreased their treatment compliance and increased the risk of iatrogenesis. This was also shown in other surveys [31]. Limited access to their HCP jeopardized patient adherence to treatment, potentially resulting in the long-term worsening of their chronic condition and contributing to the high burden of NCDs [23]. The annual costs of medication non-adherence have been estimated to be as high as EUR 1.25 billion in Europe and USD 290 billion in the USA [32]. Additionally, many patients had to cope with the fact that many medical surgeries and consultations were postponed or canceled (52% during the first lockdown in our study). Soon after the start of the pandemic, triaging was the most widely adopted strategy worldwide to ensure high risk NCD patients continuing access to critical monitoring services and treatments without increasing the risk of COVID-19 contagion at hospitals [33]. For this purpose, the COVID-19 patient management algorithm published by the WHO might be applicable to NCD patients [34]. Months after the start of the pandemic, the frequency of physician consultations was still reduced, and patients had to adapt their consultation mode to get appropriate information on their chronic diseases. Therefore, more patients started using telemedicine (advice by telephone or online means) for their medical visits [35][36][37][38]. Telemedicine might be an additional and substitutive way of providing continuing essential healthcare services. However, as the use of telemedicine is highly dependent on the availability of technology and expertise, this technology might not be accessible to the elderly population; guidance and support are therefore urgently needed [10]. Health specialists and pharmacists might also be involved in the development of the telemedicine especially [39]; due to the pandemic in the past year, the rapid growth of telemedicine has renewed the debate of physician licensing in many countries [40]. In that respect, this pandemic might have been an opportunity to learn how to leverage healthcare technologies [28]. To better cope with their chronic diseases during the pandemic, patients needed more specific information, especially on how COVID-19 infection might impact their current pathology in terms of the associated risks and appropriate care management during lockdown. At the beginning of the pandemic especially, patients had difficulties distinguishing reliable medical facts and misinformation among the huge amount of available information. During the pandemic, governmental agencies such as the WHO were providing relevant information on patients with NCDs and the COVID-19 pandemic [10,11]. In our study, while television was the most used source of information during the pandemic, it was not very well trusted; on the other hand, if physicians were the most trusted source of information by patients on their chronic disease, they were not consulted enough. In addition to physicians, different sources with validated information, such as the pharmacist or patient advocacy groups, related to specific pathologies were available during the lockdowns. As rehabilitation care services as well as health and social care at home were the most commonly disrupted services worldwide [12], NCD patients were essentially relying on non-professional carers/caregivers-most of them being family members-for daily care, and in some cases, including technical medical care as well as for keeping physicians updated on the patient's condition. Due to the pandemic, the disruption of care services dangerously aggravated a pre-existing situation and led to an overreliance on carers for long-term care, jeopardizing their situation [41]. Considering the number of people living in the same household as the respondents and the deterioration reported by the latter concerning their health and well-being, family members living under the same roof were more likely to be impacted by this acute situation through additional stress and informal care provisions. The lack of awareness of the role of the informal carer/caregiver-often taken for granted by the carer as well as by the person in need of care-was most likely to explain the fact that only 10% of respondents mentioned explicitly having asked an informal carer for help. Most informal carers/caregivers had to cancel or reschedule their own medical consultations or surgeries to manage the health care of their close relatives with NCDs during the pandemic and reported having experienced more problems in terms of reconciling work and care since the start of the coronavirus pandemic [42][43][44]; this impact of the pandemic on informal carers/caregivers-on top of their daily burden in non-pandemic times [45]-has been largely overlooked. Both patients and informal carers/caregivers suddenly had increased feelings of isolation and were in urgent need for information to educate themselves on rehabilitation procedures. During the pandemic in France, caregiver associations developed free-accessible online dedicated videos and print documents to train chronic patients and their non-professional/informal carers/caregivers on how to manage their essential care and condition in different life situations [46]. Emerging situations like the COVID-19 pandemic especially revealed that such educational programs are crucial for patients and informal carers to lower their anxiety and stress level, therefore reducing the risk of comorbidity development. In an assessment of the situation performed by the WHO in patients living with NCDs, several countries worldwide requested explicit support in developing communication materials addressing NCDs and their risk factors in the context of the pandemic [10]. This pandemic therefore stressed the long-needed support of informal carers/caregivers and the need for integrated care supported by the WHO and the European healthcare system [47,48]. Patients with NCDs, especially those presenting with severe conditions and/or multimorbidities, rely on integrated care programs; those were greatly affected during the pandemic as healthcare services were very busy managing the COVID-19 crisis. This might have had devastating consequences, particularly among NCD patients requiring regular symptom monitoring and the adjustment of complex drug regimens [47]. Although the care management of patients living with NCDs is mainly a patient-centered approach, integrated care including all of the relevant primary to tertiary health services is a key factor, especially for patients with more severe pathologies and multimorbidities, who are more disabled. In fact, patients with NCDs might be described as living in an ecosystem, in which central informal carers/caregivers play a key role, and therefore, are dependent on all of the surrounding health care management services; the coordination of care among the different health care partners facilitates the appropriate and timely delivery of health care services to NCD patients [49]. Integrated care means HCPs, secondary/hospital care specialists, pharmacists, rehabilitation specialists, residential care specialists, mental health specialists, informal carers/caregivers as well as patient and caregiver associations work in collaboration to improve care management for patients with NCDs. This point raises the need for a health care system reform to implement or strengthen NCD integrated care management programs, especially in terms of linking hospital and private practices or implementing care managers, such as in the UK, for the organization of a patient's daily life after hospital discharge. Governments around the world need to understand how to support health services, align on clear and concise healthcare messaging from a globally established reliable source of information, and have emergency plans in place to mitigate the impact of any future sanitary crises [10,50,51]. Differences in the health care systems could also explain the differences observed. Health coverage in Europe is universal. Having different structures of interactions between insurers, providers, and patients, all European healthcare systems aim to provide care to everyone on the grounds of free access, equality and equity, and fairness: no matter how much patients earn, patients receive a basic package. Unlike Europe, not all American citizens have access to publicly-funded insurance; government funds are available for certain Native American tribes, military families, and veterans. There is also national health insurance, called Medicare, which covers people over 65 and some people with disability status, people with end-stage renal disease, and amyotrophic lateral sclerosis [52]. These differences could especially explain the impact in the access to HCPs or on the social isolation/anxiety that have been described more often in the US patients. Additionally, disparities in terms of satisfaction towards the information received could be mostly due to the differences in the health care systems and reimbursement access between Europe and the USA. This study presents several limitations, mostly inherent to its design. The following bias should be mentioned: selection bias (underrepresentation of older patients on Carenity communities or patients without online access or high-IT and literacy bias, overrepresentation of patients worrying about their health status, and members of Carenity communities), self-reported perceptions and symptoms, and recall bias, especially concerning the retrospective aspect of the study (wave 1). However, as the answering of the questionnaire was done directly by the patients in absence of their physician, the desirability bias was greatly limited. Moreover, the high similarity of the results obtained in both waves validates our study concept. Finally, belonging to a patient community likely reflects the need to self-assess and compare the levels of the disease. Poorly controlled, highly symptomatic, and heavily treated patients may then be overrepresented, and our findings should be extrapolated only cautiously. However, the majority of people use the Internet. In 2019, it was estimated that 86.7 percent of people living in developed countries used the internet [53]. Moreover, it has been shown that Carenity communities, when compared with the SNIIRAM database (a quantified claim database available in France where all individual refunding acts are listed but without any disease-related information), reflect the main characteristics of online users willing to share their experience with a disease, with an over-representation of female patients aged from 25 to 54 [54]. The study population was based on patients registered on the Carenity platform from Europe and USA and did not include other relevant sampling procedures and methods. Despite the study recruitment comprising of a homogeneous population of patients through the Carenity platform, no generalizations of the study can be assumed. --- Conclusions During the pandemic, healthcare services disruptions coupled with newly adopted negative behaviors put patients living with NCDs at high risk for the worsening of their chronic condition or developing multimorbidities. The burden of NCDs, which increased during the pandemic, has recurrent health and economic consequences for the patients' ecosystems. Hopefully, lessons from the COVID-19 pandemic will be learned, and emergency plans will be put in place to protect the NCD vulnerable population against upcoming outbreaks. NCDs were included for the first time in the United Nations 2030 agenda for sustainable development as a health priority to reinforce WHO engagement by 2030; the aim is to reduce premature mortality from NCDs by one third through prevention and treatment as well as to promote mental health and well-being [55]. Our study highlights the urgent need to restore a patient centric approach to health care-which has been clearly endangered by the COVID-19 pandemic-and to empower patients affected by non-communicable diseases (NCDs). --- Data Availability Statement: The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy laws. --- Author Contributions: Conceptualization, C.P., E.P., M.C., R.P., J.v.V., K.G. and A.-F.L.-P.; data curation, E.P.; formal analysis, E.P.; methodology, C.P., E.P., M.C., R.P., J.v.V., K.G. and A.-F.L.-P.; project administration, E.P.; software, E.P.; supervision, C.P., M.C., R.P., J.v.V., K.G. --- Institutional Review Board Statement: The study was conducted according to the guidelines of the Declaration of Helsinki. Ethical review and approval were waived for this study, as the aim of the study is to determine the patients' insights from their experiences, and informed consent was given by all patients prior to answering the survey. Informed Consent Statement: Informed consent was obtained from all patients involved in the study. stock holdings from Pfizer, Viatris, BMS, GSK, Gilead, and AstraZeneca. AFLP is a Quality Manager at Pfizer PFE France, a Viatris company.

An international online patient community, Carenity, conducted a patient study in two independent waves among adults affected by non-communicable diseases (NCDs) in Europe and in the United States of America (USA). The study aimed to assess the real time impact of the coronavirus disease 2019 (COVID-19) on the medical conditions of patients with NCDs, their access to health care, and their adaptation to daily life as well as to describe their sources of information on COVID-19 and their needs for specific information and support. During the pandemic, 50% of the patients reported a worsening of their medical condition, and 17% developed a new disease. Additionally, 26% of the respondents reported an impact of the pandemic on regular/long-term treatment intake. 54% of the patients felt very or completely socially isolated and reported a strong impact of the COVID-19 pandemic on their stress level and state of mind, with higher levels observed in the USA compared to Europe. 59% of the respondents wished to have received additional information regarding the risks associated to their medical condition during the pandemic. Television was the most used source of information, whereas physicians were the most trusted one. This study describes the substantial impact of the COVID-19 pandemic on NCD patients.

populations (6). Hispanic patients with higher depressive symptoms are 3.27 times more likely to have poor glycemic control than Hispanics with lesser depressive symptoms (4). A study with 1,293 patients showed that diabetes-related complications, symptoms, and distress were correlated with emotional burden (5). Many scholars regard depression as having a bidirectional relationship with diabetes management and distress (6,7). The social context in which patients with diabetes are embedded is a significant factor for self-care behaviors (8). Social support has been mostly studied using constructs measuring the quality of the social context in association with diabetes management (9,10). Social support can affect why and how individuals manage illnesses, help patients believe that they have the efficacy or capacity to implement recommended self-care behaviors, and reveal solutions to coping with barriers impeding successful diabetes management (11). A meta-analysis concluded that social support is effective in modifying behavior for better diabetes management (12). Older adults with diabetes who reported that they received support with taking medicine, engaging in enough physical activity, and going to health care providers were more likely to have improved health after 7 years (9). The patient-centered medical home (PCMH) model is focused on treating comorbid depression and diabetes (13). Psychosocial interventions have recognized the significance of social support for low-income Hispanic patients with comorbid depression (14,15). Social support was recognized early as a promising target for psychosocial services developed to treat depressive symptoms among patients with chronic illnesses (16). In some models, promotoras and community health workers aim to improve diabetes management by enhancing social support and social resources, and strengthening social support is one of the main therapeutic activities shown to increase perceived social support (15). A study focusing on Hispanic patients with cancer with comorbid depression found that social support was a significant mediator that explained differences in daily functioning among patients with cancer who received collaborative depression care at safety-net clinics (17). Yet, few empirical findings exist regarding relationships between changes in depressive symptoms and diabetes management. Although one study examined correlations between changes in depressive symptoms and frequency of four self-care behaviorsdhealthy diet, exercise, blood glucose monitoring, and foot caredfindings were mixed (18). Because adherence to diabetes management can be measured in many ways (19)(20)(21), further studies are needed. To our knowledge, no study has examined whether improved social support can explain changes in diabetes management among individuals following a patient-centered care regimen, which is becoming the norm due to the implementation of the Affordable Care Act (22). Although social support has been identified as a significant predictor of diabetes management in observational studies (9,11,12), this correlation was not found in a study involving less acculturated Hispanic patients with diabetes (23). Specifically, perceived social support was not correlated with adherence to self-care behaviors among lowincome Hispanics living in Texas, even when perceived support was very specifically measured in terms of behaviors associated with diabetes management (23). Thus, studies are needed to empirically examine whether changes in social support observed in primary care can explain the variance in diabetes management in this population. Last, depressive symptoms and perceived social support are often correlated, and their associations with diabetes management may not be clearly observed unless they are examined in the same statistical model. Patients with higher depressive symptoms also tend to present with increased interpersonal conflicts and lower social support (24,25). A prospective study found patients with type 1 diabetes with higher depressive symptoms had significantly lower social support (26). By examining changes in depressive symptoms and social support, we can document the effect sizes of those two variables in explaining the variance in diabetes management. Tofill these researchgaps, we examined whether changes in depressive symptoms and social support during 6 months postbaseline can explain self-efficacy and adherence to recommended self-care behaviors for patients with diabetes at the 6-month follow-up. We conducted secondary analyses of data collected in a randomized clinical trial (RCT) that compared the effectiveness of a promotora-led self-management program and a newly launched PCMH model by a local government regarding diabetes management among Hispanic patients with diabetes relying on safety-net clinics in a low-income community. A promotora is a community health worker for Hispanic patients who continues to practice indigenous customs and conform to cultural norms (15). By providing culturally tailored services boosting diabetes management, this intervention intends to alleviate emotional distress as well as enhance diabetes management (15). Hypothesis 1: changes in depressive symptoms and social support during 6 months postbaseline will independently predict differences in self-efficacy and adherence to recommended self-care behaviors for patients with diabetes at the 6-and 12-month follow-ups. --- RESEARCH DESIGN AND METHODS --- Sampling and Procedures The study analyzed data collected from Hispanic patients with diabetes screened with comorbid depression at baseline in the A Helping Hand (AHH) study. AHH recruited Hispanic patients diagnosed with diabetes, cardiovascular disease, or heart failure from April 2014 to May 2015. In total, 348 patients from three safety-net clinics in East Los Angeles, CA, met eligibility criteria and completed informed consent and a baseline survey. We analyzed data from patients previously diagnosed with diabetes, and 333 patients (95.7%) met this eligibility. Initial analyses determined whether changes in depressive symptoms and social support predicted diabetes management at the 6-month follow-up. We found that 82 patients (24.6%) did not complete the 6-month survey, creating a sample of 251 patients (75.4%) for analysis. We conducted further analyses to determine whether changes in depressive symptoms and social support predicted diabetes management at the 12-month follow-up. Of 333 patients with diabetes, 113 patients (33.9%) did not complete either the 6-or 12-month follow-up, leaving 220 patients (66.1%) for analysis. AHH was an RCT that examined the effectiveness of a promotora-led self-management support program compared with a standard PCMH team featuring a physician, nurse, and medical assistant (15). Promotoras provided psychosocial services focused on six areas: engagement for initial rapport building, problem formulation to list targeted problems, education for teaching self-care management strategies and health information, action planning to develop action steps and implementation, community resource navigation and referrals to providers, and evaluation through feedback (15). Promotoras primarily provided services during in-person meetings, although services were delivered by phone when a patient was not available for an in-person meeting. In addition to six main sessions, three monthly booster sessions were offered to activate self-care management (15). PCMH usual care included depression screening by a primary care provider, referral to community health workers who were newly hired clinic promotoras, referral to behavioral health specialty care, and activation of patient caremanagement (15). Themain outcome study found no significant differences in terms of depression outcomes, psychologicalhealth,self-care management,physicalcondition,stressandsocialsupport,and glycemiccontrolbetweenthepromotora-led intervention and PCMH usual care (15). The AHH study was reviewed and approved by the institutional review board of an affiliated university. Table 1 features descriptive statistics of demographic and clinical factors in the total sample and among patients who were analyzed or dropped out at the 6-month follow-up. Among the baseline and 6-and 12-month follow-up surveys gathered from the AHH, our study analyzed data from patients who completed baseline and 6-month follow-up surveys (n = 251; 75.4%); 82 patients (24.6%) did not complete the 6-month survey. We compared the means of demographic and clinical factors between those whose data were analyzed and those who had dropped out. No demographic and clinical factors at baseline differed significantly between these two groups other than marital status (P, 0.05). --- Measures Depressive symptoms were assessed via the nine-item Patient Health Questionnaire (PHQ-9). Depressive symptoms are a continuous variable ranging between 0 and 27. The PHQ-9 has adequate internal consistency and test-retest reliability (27). It is widely used as a screening tool to discriminate between racially diverse primary care patients with and without depression (28). The internal consistency of the PHQ-9 measured at baseline and 6-month follow-up was 0.53 and 0.88, respectively (N = 251). To find a reason for lower internal consistency at baseline, we conducted a descriptive analysis on distribution of responses to individual items ofthePHQ-9. Therewasmorevariabilityof responses to each item at baseline when compared with the 6-month follow-up, andthisvariabilityseemedtobetranslated into lower internal consistency at baseline (Supplementary Data). Overall, more patients endorsed questions on "little interest," "feeling down," "trouble falling or staying asleep," and "feeling tired." Changesindepressivesymptomswerecalculated by comparing depressive symptom scores at baseline and 6-month follow-up. Higher values represent higher depressive symptoms at 6-month follow-up compared with baseline. In other words, patients with positive values presented higher depressive symptoms at the 6-month follow-up compared with baseline. Perceived social support was measured with the eight-item Modified Medical Outcomes Study (MOS) Social Support Survey (29). This abbreviated version was extracted from a 19-item survey that was part of the MOS in the early 1990s (30). A psychometric study of the modified version found comparable internal consistency, convergent validity, and divergent validity when compared with the longer version (29). The same study presented evidence for factorial validity; namely, that the instrument has two factors: instrumental social support (items 1-4) and emotional social support (items 5-8). Items ask how often someone is available when a respondent is in a particular situation, such as being confined to bed and needing to visit a doctor. Response options include: 1 (none of the time), 2 (a little of the time), 3 (some of the time), 4 (most of (20,21). Response options range between 1 (none of the time) and 5 (most of the time). The MOS-SAR was developed to measure chronic illnesses, including diabetes, hypertension, and heart disease, and 11 selfcare behaviors are included for final score calculation. They include taking prescribed medication, following a low-fat or weight-loss diabetes diet, exercising regularly, eliminating or reducing smoking and alcohol use, limiting stress, checking feet regularly, checking blood glucose, carrying supplies needed for self-care, and carrying something with sugar. An average was calculated based on responses to a second set of questions regarding frequency of implementing self-care behaviors recommended by physicians. The internal consistency of the MOS-SAR at baseline and 6-month follow-up was 0.83 and 0.65, respectively (N = 251). Self-efficacy related to diabetes management was measured with the Self-Efficacy for Managing Chronic Disease scale (31). This measurement has six questions that ask about the extent to which respondents are confident about keeping fatigue, pain or physical discomfort, emotional distress, or other symptoms from interfering with what they want to do, being able to do things that can reduce the need to see a doctor, or how much illness affects their everyday life. Response options range between 1 (not at all confident) and 10 (totally confident). Average scores for these six items were calculated. Psychometric properties were empirically examined in a study featuringeight samples fromfive countries, including studies conducted in Spanish and English (32). In our study, internal consistency of this measure at baseline and 6-month follow-up was 0.94 and 0.93, respectively. --- Analytic Strategies We examined longitudinal patterns of the variables of interest and conducted a series of paired t tests comparing the means of data at baseline and 6-and 12month follow-ups. Then we developed hierarchical regression models to examine correlations between changes in depressive symptoms and social support variables during 6 months postbaseline and self-efficacy and adherence to recommended self-care behaviors at 6-and 12-month follow-ups. Each hierarchical regression model contained four submodels. Model 1 only examined correlations between the outcome measured at the 6-and 12-month follow-ups and the baseline value of the outcome while controlling for potential confounding factors, including health status when the outcome was measured, study arm, diabetes types, age, sex, education, language, and dysthymia (15). By controlling for the baseline value of the outcome, further analyses can rule out innately different levels of self-efficacy and adherence to diabetes management at baseline. By adding changes in depressive symptoms, model 2 estimated coefficients without the social support variable. Model 3 estimated coefficients for change in total social support while considering changes in depressive symptoms, together with other covariates included in model 1. Due to multicollinearity, we had to create model 4, which included changes in instrumental social support and emotional social support, changes in depressive symptoms, and other potential confounding variables. As the post hoc analyses, we controlled for the self-efficacy and adherence to diabetes management at the 6-month follow-up when examining diabetes management at the 12-month follow-up. These post hoc analyses were intended to find the time at which changes in depressive symptoms and social support affected diabetes management. SPSS 24.0 (SPSS Inc.) was used for statistical analyses. For hypothesis testing, we used P, 0.05 to signify statistical significance. --- RESULTS Table 2 displays longitudinal patterns of the means and SDs of depressive symptoms, perceived social support (including total, instrumental, and emotional support), self-efficacy and adherence to diabetes management, and the eight individual selfcare behaviors in the MOS-SAR. Mean change in depressive symptoms indicates change between baseline and 6-month follow-up. Participants had an average decline of 7.21 in depressive symptom scores between baseline and 6-month follow-up. Social support generally improved; for instance, total social support increased by 21.43%. Changes in selfefficacy and adherence to diabetes management during the same period were relatively limited compared with changes in depressive symptoms andsocial support. With regard to the eight individual self-care behaviors, which are recommended for patients with diabetes, patients reported lower adherence to regular exercise, as reflected by a lower mean compared with other behaviors across the three waves. Table 3 presents results from hierarchical regression models examining diabetes management outcomes at the 6-month follow-up. This hierarchical regression examined four models in a consecutive manner. Two sets of hierarchical regression analyses were conducted with self-efficacy and adherence to diabetes management as the outcome. Changes in depressive symptoms during 6 months postbaseline were significantly correlated withself-efficacyatthe6-monthfollow-up (P, 0.001). This suggests that patients with more decline in depressive symptoms over 6 months reported higher selfefficacy to manage diabetes. Effect sizes of changes in depressive symptoms were ;0.40, meaning that when patients with diabetes experienced a decline in depressive symptoms of 1 SD, they increased self-efficacy by 0.40 SD. Adding changes in depressive symptoms (model 2) accounted for 15.0% of the adjusted explained variance. In model 3, changes in total social support were significantly correlated with self-efficacy (P, 0.01), suggesting that patients whose total social support increased over 6 months reported higher self-efficacy at the 6-month followup. Adding total social support change increased the adjusted explained variance compared with model 2 by 2.5%. Results from model 4 show that changes in emotional social support were correlated with the outcome (P, 0.05), yet instrumental social support was not (P. 0.05). With respect to adherence to diabetes management, changes in depressive symptoms over 6 months significantly predicted differences in adherence to diabetes management in models 2, 3, and 4 (P, 0.001). The direction of the association indicates that patients reporting lower depressive symptoms at the 6-month follow-up engaged in significantly more frequent selfcare behaviors at the 6-month follow-up when the baseline value of the outcome was controlled for. Adding changes in depressive symptoms (model 2) increased the adjusted explained variance by 7.1%. Yet, no social support variables were significantly correlated with adherence to diabetes management (P. 0.05). Table 4 presents results from a hierarchical regression model examining diabetes management outcomes at the 12-month follow-up. Changes in depressive symptoms explained different levels of self-efficacy (P, 0.01). Standardized coefficients were reduced by almost half compared with analyses focusing on the 6-month follow-up. Changes in depressive symptoms (model 2) added 2.7% to the explained variance compared with model 1. No social support variables were significantly correlated with self-efficacy related to diabetes management at 12-month follow-up. When adherence to diabetes management at 12-month follow-up was examined, changes in depressive symptoms were still significantly correlated with the outcome in model 2 (P, 0.05), model 3 (P, 0.05), and model 4 (P, 0.01). Standard coefficients reflecting changes in depressive symptoms declined to a lesser extent compared with those for self-efficacy related to diabetes management. Social support variables were not significantly predictive of diabetes management at the 12-month follow-up (P. 0.05). Post hoc analysis was conducted controlling for the self-efficacy and adherence to diabetes management at the 6-month follow-up when examining diabetes management at the 12-month follow-up. We found that changes in depressive symptoms were not a significant factor in this post hoc analysis (P. 0.05). This post hoc finding suggests that the influence of changes in depressive symptoms on self-efficacy and adherence to diabetes management --- CONCLUSIONS The AHH study offered a suitable setting to examine this research question because one group received PCMH usual care and the other group received promotora-led selfmanagement services. Due to services offered in the PCMH model, as designed and implemented by the Los Angeles County Department of Health Services, the primary outcome study found no difference between the two groups in depression, diabetes management, physical health, and glycemic control (15). Changes in depressive symptoms during 6 months postbaseline consistently predicted self-efficacy and adherence to diabetes management. Results regarding relationships between social support and diabetes management were mixed. Patients reporting an increase in total social supportandemotionalsocialsupportduring 6 months postbaseline reported higher selfefficacy at the 6-monthfollow-up. Yet,social support variables were not correlated with self-efficacy at the 12-month follow-up or adherence to diabetes management at 6-or 12-month follow-ups. Consistent empirical findings have supported the significance of mental health care for patients with chronic illnesses, including diabetes (33,34). A study found that medication adherence and glucose monitoring were clearly enhanced when participants received cognitive behavioral therapy (35). However, few studies have attempted to assess whether changes in depressive symptoms observed in primary care explain the variance in diabetes and chronic illnesses. A systematic review concluded that no intervention studies focused on comorbid depression had presented evidence of a statistically significant relationship between depression and self-care behaviors (33). Also, some studies showed that adherence to selfcare behaviors was not significantly higher in the intervention group when compared with the control group, although the intervention group demonstrated significantly lower depressive symptoms on average in a clinical trial (34,36). An exceptional study that analyzed data collected from an RCT testing the effectiveness of culturally tailored collaborative depression care for low-income Hispanic patients found that only exercise was prospectively predicted by depression remission observed at 12 months postbaseline (18). Yet, this particular study measured adherence to diabetes management with the Summary of Diabetes Self-Care Activities, which does not suggest aggregating scores for four individual self-care behaviors (19). In other words, this study was not able to test whether a decline in depressive symptoms observed in primary care predicts enhanced overall adherence to diabetes management (18). Our study measured adherence to diabetes management with the MOS-SAR, allowing us to examine the effect of changes in depressive symptoms on adherence to overall self-care behaviors. Despite our findings supporting the significance of treating comorbid depression to selfefficacy and adherence to diabetes management, our study is unable to provide evidence that promotora-led intervention can enhance diabetes management by virtue of treated depressive symptoms. Trief et al. (37) conducted mediation analysis testing significance of self-care behaviors in associations between receiving an intervention and HbA 1c level. The same analytic approach would offer more convincing evidence for the argument that alleviation of depressive symptoms is necessary for improved diabetes management, yet we were not able to use the same analysis because the parent RCT failed to find different levels of diabetes management in the promotora-led intervention group when compared with the PCMH model control group (15). Further studies should examine significance of depressive symptoms as a mediator that explains different diabetes management levels between an intervention group and a control group in an RCT. Contrary to predictions based on the existing literature (9,11,12), changes in social support had a negligible association with both self-efficacy and adherence to diabetes management. We believe that three possible explanations deserve furtherconsideration.First,asourstudyfound, the strength of the association between changes in social support and diabetes management could be too negligible to detect in a low-powered study. A review study noted different findings from observational studies depending on sample size (10). Studies with larger samples tend to find significant relationships between social support and diabetes management (10). Yet, this reason does not seem convincing because effect sizes of social support-related variables were minimal in our study. Second, it is possible that social support only works as a buffer to the negativities from social and diabetesrelated stressors (38). For instance, Baek et al. (39) found that social support was associated with diabetes-related distress only if patients had higher diabetesrelated burden. This finding suggests that social support might become potent only if patients experience more diabetes-related burden, and further studies are recommended for empirical scrutiny for this hypothesis. However, additional analyses testing moderating effects of social support did not offer evidence for buffering effects when functional disability was used as a proxy for diabetes burden (40). Lastly, it is possible that measuring social support may not capture nuanced mechanisms by which social support facilitates selfefficacy and behavioral modification regarding chronic illness management. For instance, a study analyzed data from 463 adults and found that significant correlations between social support from family, friends, and community and individual self-care behaviors were observed only if social support was specific to behaviors (41). In other words, rather than perceptions of general social support associated with chronic illnesses as measured by the eight-item Modified MOS Social Support Survey, supportive behaviors specifically targeting individual self-care behaviors, such as healthy diet or exercise, are likely more significant predictors for diabetes. Last, social support may not be the sole function of social networks (41)(42)(43). Social support is one of many psychosocial mechanisms by which diabetes management is affected by input from social networks (41). This effect is often more pronounced among immigrants or people living in ethnic enclaves, which wasgenerallythe caseinour sample (44). The social networks in which lowincome Hispanic individuals are embedded may influence diabetes management outcomes by virtue of different levels of social support, availability of accurate health information, and norms and rules shared among in-group members. Functions other thansocialsupportlikelyinfluencewhether Hispanic patients with diabetes adhere to a primary care physician's recommendations (23). Ourstudyhasseverallimitationsspanning different research activities. First, the AHH study conducted recruitment using convenience sampling at three safety-net clinics that were not randomly selected. Findings from the study should not be generalized to the population of interest (i.e., patients with diabetes and elevated depressive symptoms who use patient-centered primary care safety-net clinics). To our knowledge, no population-level research had previously studied this particular population. Second, ;25 and 34% of patients did not complete 6-and 12-month follow-up surveys, respectively, and these patients were excluded from the study. We found no significant baseline differences in most demographic, clinical, and psychosocial attributes other than marital status between those who remained and dropped out. However, it is possible that patients dropped out because they might not have benefited from either promotora-led selfmanagement services or PCMH usual care to the same extent as those who remained. Our study presented empirical results explicating whether changes in depressive symptoms and social support can explain differences in self-efficacy and adherence to diabetes management among lowincome Hispanic patients with diabetes and elevated depressive symptoms receiving patient-centered primary care from a safety-net clinic. We found evidence of thesignificantassociationsbetweenchanges in depressive symptoms and self-efficacy and adherence to diabetes management. Further studies using different methodological approaches to measure different functions of social networks (e.g., informationdisseminationandnormativepressure) are highly recommended.

This study examined whether changes in depressive symptoms and social support prospectively predicted diabetes management among Hispanic patients with probable depression in patient-centered medical homes at safety-net clinics in East Los Angeles, CA.Data were collected from 251 patients enrolled in a randomized clinical trial testing the effectiveness of a self-management intervention assisted by a promotora (Hispanic communityhealthworker).Cross-laggedanalysesexaminedassociationsbetweenchanges in depression symptoms and social support between baseline and 6-month follow-up and self-efficacy and adherence to diabetes management at the 6-and 12-month follow-ups.Changes in depressive symptoms predicted self-efficacy and level of adherence at the 6-and 12-month follow-ups. Changes in total social support and emotional social support were correlated only with self-efficacy regarding diabetes management at 6-month follow-up.Decline in depressive symptoms is a reliable predictor of improvement in self-efficacy and adherence to diabetes management. Further studies are recommended to study psychosocial mechanisms related to social relationships other than social support that affect diabetes management. Hispanics with diabetes are less likely to follow recommended self-management instructions from health professionals, and this lower adherence to self-care behaviors is believed to explain racial disparities in glycemic control (1). Hispanic older adults with diabetes have severe health consequences, such as increased risk of kidney disorder and mortality, as noted in a 7-year study (2). Observed disparities in diabetes management appear to be related to social stressors, such as financial strain, and lack of resources for self-care behaviors (3). Studies have well documented the effects of comorbid depression on diabetes management among Hispanics with diabetes (4,5) as well as diverse

Introduction In the past, autonomous vehicles (AVs) were presented as the solution to all problems related to transportation. In fact, they were announced as vectors for more sustainable mobility, considering the broader meaning of this term. As known, the adjective "sustainable" implies ( i) environmentally friendly, ( ii) useful for society, and ( iii) economically viable. According to past (or interested) statements, AVs would eliminate congestion, thus contributing to travel time savings, more productivity, and fewer emissions (note that, even disregarding the fact that AVs will mostly be electric, emissions increase in congestion and stop-and-go episodes), and they would also drastically reduce the number of accidents. If this were true, we could indeed say that they are sustainable. However, over time, many doubts have arisen regarding the above statements. First, it is quite doubtful if AVs in the sense of fully automated vehicles (SAE 5 level; [1]) will one day become reality. Second, several studies have shown through simulation that the introduction of these vehicles would in fact imply an increase in the current congestion levels [2,3]. Finally, various surveys and studies have shown that some users are fascinated with the idea of AVs and would like to adopt them, while others would fear to travel in or near these vehicles. Note that AVs will have to coexist with less automated vehicles (and their drivers) for a long time. However, there is positive news. For example, it seems quite clear that vehicles with automation levels 3 to 4 would suffice to change the mobility system, as passengers would only have to resume the driving task on specific occasions, and these levels of automation are feasible. However, note that the condition for these vehicles to be able to drive efficiently (i.e., not conservatively) is that they are connected, that is, that they turn out to be connected autonomous (or automated) vehicles (CAVs). These vehicles will 2 of 22 be able to use, among others, vehicle-to-vehicle (V2V) communications and vehicle-toinfrastructure (V2I) communications to exchange information among themselves and with the infrastructure. With appropriate ad hoc traffic management strategies taking advantage of these data, the vehicles will be able to drive at high speeds and close distances, thus contributing to capacity increase without impairing safety [4,5]. In addition, the idea of a public transportation system based on AVs has arisen. A very efficient public transportation system taking advantage of AVs (i.e., CAVs) could foster a very positive mobility paradigm. Nevertheless, the solution to the third abovementioned issue, i.e., to AV (or CAV) rate of adoption, is still not clear. Several studies have been performed to try to understand which user profiles are less likely to adopt AVs and the reasons behind this fact. Some trends are envisaged, but there are still no clear answers. In this context, the objective of this paper is to deepen the search for the factors that determine a higher or lower level of acceptance of AVs by specific population groups, but also by the general population. This search is performed for two different cases: first, for AVs substituting current private vehicles, and second, for AVs being the basis for public transportation systems (e.g., connected autonomous buses or shuttles). With this aim, data gathered through an online survey have been subjected to exploratory analysis, ordinal regression, and exploratory factor analysis. As it will be explained, safety perceptions and an a priori AV appeal seem to be the most determining factors in AV adoption. The remainder of the paper is structured as follows: First, Section 2 summarizes the state of the art on AV adoption/acceptance; Section 3 explains the methodology used in this piece of research; Section 4 presents and discusses the main results obtained. Finally, the most significant conclusions of the paper are included in Section 5. --- Background Existing research on AV adoption can be restricted to a specific form of AV implementation or generalized, i.e., applicable to private AVs, shared AVs (SAVs), or mass transportation systems based on AVs. Similarly, some studies focus on the impact of specific factors on AV acceptance, while others are more comprehensive. The choice of the target users also implies a research decision, as their profiles can be varied or particular. In the next paragraphs, we include the main findings of highly cited works on the topic, belonging to all the above mentioned topics. For example, ref. [6] performed a literature review on AV acceptance especially focused on users' trust towards these vehicles. They concluded that, even though trust level varies depending on users' profiles, it also does according to their ethical considerations. The authors suggested that ethical issues should be agreed on, clarified, and widespread to unblock user reluctance. Ref. [7] analyzed the relationship between AV acceptance and risk preference. First, they defined three innovative parameters, one as a measure of psychometric risk preference and two economic risk preference parameters. Then, they disseminated a stated preference survey in Singapore, answered by 1142 individuals, and analyzed the association of the former parameters with socioeconomic factors. Finally, they assessed the influence of risk preference on AV choice. The results showed that all the parameters contributed to predicting the level of AV acceptance. Females, the elderly, and low-income individuals were shown to be more risk-averse and, thus, less keen on adopting AVs. However, the authors warned that risk preference had been measured and modeled from the point of view of individuals, i.e., in a general way, and that particularizing it for different mode choices could lead to different results. For their part, the goal of [8] was to understand the impact of environmental concerns on (electric) AV adoption. Even though AVs are primarily thought to be electric, the reality is that a significant percentage of those acceptance studies do not mention this fact. Ref. [8] did, however. They applied the Technology Acceptance Model [9] to 470 responses to an online survey distributed in China. The results showed that green perceived usefulness, perceived ease of use, and environmental concern have a positive relationship with the users' willingness to adopt AVs. Ref. [10] covered a wide range of possible affecting factors. In 2017, they gathered data from 2036 Finnish citizens between 18 and 64 years old. Their aim was to establish to what extent people were ready for AVs and what concerns they had. They found that highly educated men living in cities were more likely to adopt AVs and at an earlier stage. They also found that car unavailability was related to a positive perception of AVs. Interestingly, the majority of respondents expressed their wish for AVs to have the option of manual drive. Ref. [11], in turn, conducted a systematic review of those contributions based on the acceptance of private AVs, making a distinction between those including behavioral theories and those that did not. They identified seven factors related to behavioral theories, namely perceived ease of use, attitude, social norm, trust, perceived usefulness, perceived risk, and compatibility, which significantly impacted AV acceptance. Safety, performance-toprice value, mobility, value of travel time, symbolic value, and environmental friendliness also had an impact. One of the advantages initially offered by AVs was the fact that individuals usually unable to drive a traditional car, such as children, the elderly, or people with a disability, would be able to use them. Ref. [12] focused on the acceptance by this last group and is thus a good example of research targeting specific users. The researchers worked with a group of 444 UK participants with scarce or null walking abilities and a control group of people without impairments. The participants were interviewed about their feelings towards AVs and their responses were analyzed using, first, a structural topic modeling (STM) procedure and, second, a structural equation model (SEM). Results were very different between both groups, with the level of disability, the appeal of technology, the general level of anxiety, prior knowledge, the internal locus of control, and action orientation as the main factors affecting AV acceptance by people with walking difficulties. Ref. [13] also specified the target users, in this case millennials. The authors aimed to pay special attention to the impact of social factors and privacy concerns on AV adoption. To this end, they distributed an online questionnaire among Slovenian millennials born between the years 1981 and 2000 and studying at university. Afterwards, they analyzed the 359 received answers using SEM. The results highlighted the impact of privacy concerns together with perceived safety on general attitudes towards AVs. In addition, technological enthusiasm, social factors, perceived benefits, and facilitating qualities were shown to impact the participants' attitudes to AVs. As expected, positive attitudes to AVs, perceived safety, and facilitating conditions led to a higher level of AV acceptance. Among studies highlighting specific implementations of AVs, ref. [14] performed a literature review separately analyzing user willingness to use AVs or SAVs. Their analysis was based on an extended version of the Theory of Planned Behavior [15], as it also considered knowledge and perceived risk. The results were compared with those obtained after applying SEM to data collected in China through a survey answered by 906 individuals. According to this study, knowledge of AV technology and perceived risk turned out to be the main barrier to private AV and SAV adoption, which was also highly affected by the a priori attitude towards AVs. When focusing on private AVs, subjective norm plays an important role, while perceived behavioral control is more influential in the case of SAVs. SAV services were the target of [16]. Not only the role of safety and security perceptions but also that of factors such as the value of travel time were assessed, as these services imply detours to collect/drop off passengers. The authors used a multivariate integrated choice and latent variable approach to examine people's willingness to adopt these SAV services. The results showed that commuters accepted sharing the vehicles with strangers more willingly than users who travelled for other purposes. In any case, the detour-related travel time was shown to affect acceptance more negatively. High-income individuals were less sensitive to this fact but only if they were able to work when travelling. Ref. [17] also focused on SAVs, specifically AV taxis, i.e., fairly small vehicles. They organized a panel of 834 individuals living in the USA, who were asked several questions. An exploratory factor analysis helped to identify four attitudinal dimensions: technology acceptance, risk-taking, traffic regulation, and driving enjoyment. Then, the authors used multinomial logistic regression to assess the influence of these factors on general attitudes towards AVs and on people's willingness to adopt them, either as private vehicles or in the form of shared AV taxis. The results indicated that early adopters of technology and users with high compliance with traffic regulations had a positive attitude towards AVs and SAVs, as opposed to risk-averse people. Nevertheless, overall acceptance was found to be quite low. For their part, ref. [18] aimed at quantifying the role of the subjective value of travel time on peoples' shift to AVs, but also to traditional public transportation. They elaborated and distributed a comprehensive survey, which was answered by 2229 Northern California commuters. Then, they defined a revealed preference mode choice model accounting for the impact of multitasking attitudes and behavior on modal choice. Results showed that public transport users willing to perform productive activities while commuting would shift to the private car if these tasks were not possible. If performing these tasks were possible in public transport but also in private vehicles (i.e., AVs), the latter would gain a 1.5% share. Ref. [19] were more focused on methodologies than on results, even when they contributed with both to the state of the art. Indeed, their goal was to define a robust and generalizable methodology to assess AV acceptance independently of the type of AV service. They tested four frameworks that included sociodemographic as well as latent behavioral factors. A survey containing questions about all of them was launched online and answered by 507 people living in the Greater Dublin Area of the Republic of Ireland. The researchers used the gathered data in confirmatory factor analyses to specify and demonstrate scale reliability of indicator items, and convergent and discriminant validity of relationships among latent variables such as perceived benefits and perceived ease of use of AVs, public fears and anxieties regarding AVs, subjective norm, perceived behavioral control, attitudinal factors, etc. Then, they tested the four proposed models and validated their usefulness in assessing AV adoption intentions and choice between AV modes, namely private AVs, SAVs, or public transport based on AVs. Regarding the particular results obtained from their sample, they found that, despite the existence of fears in relation to safety and security, most people expected benefits from the implementation of AVs. The authors also found a strong positive relationship between environmental concerns and technology, being attitudes towards AV-based public transport especially favorable. As in previous studies, a priori favorable attitudes towards technology were linked to higher confidence and a feeling of being in control when using AVs. With regard to sociodemographic factors, one of the main conclusions was that females were more skeptical about technology-related benefits to society and, subsequently, more likely to adventure negative safety and security effects linked to AVs and to have lower expectations of AVs' potential positive impacts. The following studies provide interesting insights, partly because they used data beyond those obtained from stated preference surveys. Ref. [20] applied both the Theory of Planned Behavior [15] and the Technology Acceptance Model [9] to assess people's intention to shift to AVs in the form of private vehicles or AV-based public transportation solutions. Very interesting data were collected from a gender-balanced sample of 74 individuals between 25 and 64 years old, who participated in a driving simulation experiment and answered a survey afterwards. The authors performed a hierarchical regression analysis, which highlighted the role of attitude toward the behavior, subjective norms, perceived behavioral control, and ease of use as predictors of intention to use AVs. As in other studies, AVs that allow manual control were shown to be accepted more willingly. Finally, ref. [21] is one of the few contributions to the state of the art that analyzed real experiences and perceptions of AV travelers. Specifically, 197 users of driverless shuttles implemented in the City of Vantaa were interviewed and their responses were statistically analyzed. The participants reported a higher safety perception of the driverless shuttles than of conventional buses. However, the perception of in-vehicle security of the driverless vehicles was much lower, especially in women. --- Methodology The following subsections explain the methodology used for this research, which is outlined in Figure 1. conventional buses. However, the perception of in-vehicle security of the driverless vehicles was much lower, especially in women. --- Methodology The following subsections explain the methodology used for this research, which is outlined in Figure 1. --- Survey Design and Scope The survey was defined in several iterations, after appropriately combining a number of questions that we wanted to include to account for specific factors with others that inspired previous studies consulted during the research on the state of the art. In addition, the order of the questions, the particular way in which each question was expressed, as well as the response options, were also carefully reviewed to avoid bias. Questions addressed the following topics: • Sociodemographics • Travel patterns and experiences The English version of the survey can be found in Appendix A. Depending on the questions, answers could be binary (e.g., yes or no questions), categorical, or ordinal. Most of the latter were based on the Likert scale [22], the formulation of the possible answers being dependent on the particular question addressed. The last category, the willingness-to-adopt questions, included only two questions, which asked directly about the respondents' likelihood of shifting to private AVs or public transportation systems based on AVs. These questions were the basis of the dependent variables in later analyses. The total list of variables linked to the survey and finally used in subsequent analyses is included in Table 1. Note that the survey was also translated into Spanish and Catalan in order to reach the highest possible number of respondents. Unfortunately, economic and time constraints prevented the authors from producing in-person questionnaires, so the survey was distributed online. Unless this fact introduced an unavoidable bias, we tried to reach different respondent profiles by distributing the survey through very different channels (see Table 2). --- Variable Type Denomin. --- Survey Design and Scope The survey was defined in several iterations, after appropriately combining a number of questions that we wanted to include to account for specific factors with others that inspired previous studies consulted during the research on the state of the art. In addition, the order of the questions, the particular way in which each question was expressed, as well as the response options, were also carefully reviewed to avoid bias. Questions addressed the following topics: The English version of the survey can be found in Appendix A. Depending on the questions, answers could be binary (e.g., yes or no questions), categorical, or ordinal. Most of the latter were based on the Likert scale [22], the formulation of the possible answers being dependent on the particular question addressed. The last category, the willingness-to-adopt questions, included only two questions, which asked directly about the respondents' likelihood of shifting to private AVs or public transportation systems based on AVs. These questions were the basis of the dependent variables in later analyses. The total list of variables linked to the survey and finally used in subsequent analyses is included in Table 1. Note that the survey was also translated into Spanish and Catalan in order to reach the highest possible number of respondents. Unfortunately, economic and time constraints prevented the authors from producing in-person questionnaires, so the survey was distributed online. Unless this fact introduced an unavoidable bias, we tried to reach different respondent profiles by distributing the survey through very different channels (see Table 2). We reached 380 respondents, approximately 23% from Catalonia, 29% from the rest of Spain, and 48% from other European countries. During the pre-processing, the answers of 4 of them were eliminated, as they had only partially addressed the questions. Statistically speaking, this number of respondents is enough to perform the desired analyses. However, from the point of view of population representativeness, a higher number of responses and a more balanced sample would have been desirable. This fact has been considered in the discussion section when interpreting the results. --- Data Analyses As shown in Figure 1, three different analyses were performed. We started with an exploratory data analysis, whose objective was twofold. First, we wanted to assess whether the responses were balanced, especially those related to socioeconomic factors. Second, we aimed to detect any striking particularities that could be further analyzed with more robust methods. In addition to calculating the percentages of the various responses, we looked for correlations among variables. Particularly, we looked for Spearman's correlation, as it does not require normality and can be considered the non-parametric equivalent of the Pearson product-moment correlation. The second analysis used was factor analysis. Specifically, we used exploratory factor analysis (EFA), rather than confirmatory, as we had no prior knowledge of the factors underlying the assessed variables. We used the Generalized Least Squares extraction method, as it allows working with categorical and ordinal data. Although a high number of answers is desirable when doing EFA, our sample was mathematically sufficient, considering that practice requires a minimum ratio of 10 respondents per variable. The third analysis performed was ordinal regression. Again, it was selected given the ordinal nature of the dependent variables, "Willingness to shift to private AV" and "Willingness to shift to public transport based on AV". --- Results and Discussion The following subsections describe and discuss the results obtained after performing the previously mentioned analyses. --- Exploratory Data Analysis The first analysis conducted was exploratory. One of its goals was to assess the representativeness of the sample, given that the number of respondents was limited. This analysis was performed for all the questions of the survey. Figure 2 shows the frequencies of the responses to sociodemographic questions, as well as some of those related to travel patterns. Finally, the results for the two questions directly asking about the willingness to adopt private AVs or public transportation systems based on AVs are also included. As can be seen, the sample is better balanced in terms of age than in terms of gender, with men being underrepresented (34.3%). Most respondents had a medium-high level of education (85.5%) and were full-time employees (60.1%) with an average income (473%). Interestingly, we received more responses from people related to technology (e.g., studying it or working with it), computer science, engineering, or business (74.1%). This issue is related to the non-ideal data collection method and already suggests that people from other fields were probably not even interested in answering a survey about AVs. It could be also hypothesized that people from the mentioned areas are more attracted by AVs and perhaps more likely to adopt them than individuals from other areas. The sample was well balanced in terms of the respondents having children or being childless (55.9 vs. 44.1%). Most respondents lived in cities (85.9%) and had a driving license (92.8%, with different ranges of experience) and at least one private car available for the family (88.8%). Consequently, the most used mode of transportation was the private car, even when the percentage was lower (58.8%). The car was used daily in most cases (40.2%). We can again see some bias in the sample, as most respondents were frequent drivers. Again, it could be hypothesized that these respondents would be at least more likely than, e.g., usual pedestrians, to shift to AVs. In fact, this was the case for the answers to the questions regarding the adoption of AVs. More than 50% of the respondents would be keen or very keen on adopting them either as private vehicles (50.8%) or as the basis for public transportation systems (55.3%), and more than 25% were indifferent to both possibilities. This means they would not oppose a new transportation paradigm based on these vehicles. When focusing on the respondents rejecting AV adoption, the percentage is higher in the case of private vehicles (9.3% compared with 5.3% for AV-based public transportation systems). means they would not oppose a new transportation paradigm based on these vehicles. When focusing on the respondents rejecting AV adoption, the percentage is higher in the case of private vehicles (9.3% compared with 5.3% for AV-based public transportation systems). We also checked the relationships between certain variables on the basis of the main ideas drawn from the state of the art. Figures 345show the relationship maps of some of them. On these maps, nodes represent variable categories and links represent the strength of influence between them. The larger nodes and the thicker link lines, the stronger the relationships. Figure 3 shows that there is no clear relationship between previous crash experiences as drivers and the safety perception while driving. The respondents probably felt that they were not to blame for the accident, or that it was just bad luck. Figure 4 shows that people who feel safe as passengers do so whether they are travelling in a private vehicle or on public transport. A slightly higher degree of safety is also observed among those who have not been involved in a previous accident. Finally, it can be seen in Figure 5 that there is a relationship between AV appeal and trust in AVs as means of transport. However, respondents tend to be more conservative if the passengers are not themselves but their immediate family members. We also checked the relationships between certain variables on the basis of the main ideas drawn from the state of the art. Figures 345show the relationship maps of some of them. On these maps, nodes represent variable categories and links represent the strength of influence between them. The larger nodes and the thicker link lines, the stronger the relationships. Figure 3 shows that there is no clear relationship between previous crash experiences as drivers and the safety perception while driving. The respondents probably felt that they were not to blame for the accident, or that it was just bad luck. Figure 4 shows that people who feel safe as passengers do so whether they are travelling in a private vehicle or on public transport. A slightly higher degree of safety is also observed among those who have not been involved in a previous accident. Finally, it can be seen in Figure 5 that there is a relationship between AV appeal and trust in AVs as means of transport. However, respondents tend to be more conservative if the passengers are not themselves but their immediate family members. We also checked the relationships between certain variables on the basis of the main ideas drawn from the state of the art. Figures 345show the relationship maps of some of them. On these maps, nodes represent variable categories and links represent the strength of influence between them. The larger nodes and the thicker link lines, the stronger the relationships. Figure 3 shows that there is no clear relationship between previous crash experiences as drivers and the safety perception while driving. The respondents probably felt that they were not to blame for the accident, or that it was just bad luck. Figure 4 shows that people who feel safe as passengers do so whether they are travelling in a private vehicle or on public transport. A slightly higher degree of safety is also observed among those who have not been involved in a previous accident. Finally, it can be seen in Figure 5 that there is a relationship between AV appeal and trust in AVs as means of transport. However, respondents tend to be more conservative if the passengers are not themselves but their immediate family members. With regard to Spearman's correlation, only one dependent variable, in particular the willingness to shift to private AV, correlated positively with the independent variables of education level and occupation for a significance level of 0.5. However, the correlation was weak in both cases. --- Exploratory Factor Analysis As explained, exploratory factor analysis was applied to all independent variables using the Generalized Least Squares (GLS) extraction method. In traditional factor analysis, it is assumed that the errors are homoscedastic and uncorrelated, but this assumption may not hold for real-world data. GLS extends the methodology by allowing for a more flexible covariance structure of the errors, accommodating situations where the variances of errors may vary across observations, or where correlations between errors exist. By incorporating a weighting matrix based on the specified covariance structure, GLS transforms the observed variables into a space where the assumptions of homoscedasticity and uncorrelated errors are met. This transformation enables a more accurate estimation of the factor loadings and unique variances, providing a robust framework for uncovering latent factors that contribute to observed variable variations while accounting for the complexities in the data<unk>s error structure. This technique was applied iteratively, eliminating those variables whose communalities after the extraction were smaller than 0.4. Note that factors, to be considered valid, must include at least three variables, and the correlation (loading) between these variables has to be of at least 0.3. A lower correlation would indicate a weak relationship between the variables and the factor. Table 3 includes the factors obtained using Varimax rotation and including only variables with communalities higher than 0.4. Two different extractions were made: the first one accepting all factors with eigenvalues higher than one (unrestricted extraction) and the second one limiting the number of factors to three. The Kaiser-Meyer-Olkin measure of sample adequacy was acceptable, reaching 0.685, and so was Bartlett's test of sphericity, whose significance was null. With regard to Spearman's correlation, only one dependent variable, in particular the willingness to shift to private AV, correlated positively with the independent variables of education level and occupation for a significance level of 0.5. However, the correlation was weak in both cases. --- Exploratory Factor Analysis As explained, exploratory factor analysis was applied to all independent variables using the Generalized Least Squares (GLS) extraction method. In traditional factor analysis, it is assumed that the errors are homoscedastic and uncorrelated, but this assumption may not hold for real-world data. GLS extends the methodology by allowing for a more flexible covariance structure of the errors, accommodating situations where the variances of errors may vary across observations, or where correlations between errors exist. By incorporating a weighting matrix based on the specified covariance structure, GLS transforms the observed variables into a space where the assumptions of homoscedasticity and uncorrelated errors are met. This transformation enables a more accurate estimation of the factor loadings and unique variances, providing a robust framework for uncovering latent factors that contribute to observed variable variations while accounting for the complexities in the data's error structure. This technique was applied iteratively, eliminating those variables whose communalities after the extraction were smaller than 0.4. Note that factors, to be considered valid, must include at least three variables, and the correlation (loading) between these variables has to be of at least 0.3. A lower correlation would indicate a weak relationship between the variables and the factor. Table 3 includes the factors obtained using Varimax rotation and including only variables with communalities higher than 0.4. Two different extractions were made: the first one accepting all factors with eigenvalues higher than one (unrestricted extraction) and the second one limiting the number of factors to three. The Kaiser-Meyer-Olkin measure of sample adequacy was acceptable, reaching 0.685, and so was Bartlett's test of sphericity, whose significance was null. The usual exploratory factor analysis, i.e., without an a priori restriction of the number of factors, was not conclusive enough. Of the four resulting factors, two did not reach the minimum acceptable number of related variables, i.e., three. Additionally, factor 1 and factor 4 seemed to have similar meanings. Better, if not ideal, were the results obtained from the restricted model, in which the number of factors was set to 3. The first factor is positively measured by the variables of driving license, driving experience, driving frequency, driving skills, and driver profile, while driving after drinking alcohol and safety perception while driving are negatively related to it. This factor could be interpreted as "driver features". The second factor is positively measured by age and driving experience, and negatively by preferred mode of transportation. It could be interpreted as "driving maturity". Finally, the third factor is positively explained by the variables of AV appeal and safety perception in AVs. Even when it could be seen as a degree of trust in AVs, the fact that it is only related to two variables casts doubts on its significance, as does the correlation of these variables. --- Regression Analysis Several attempts were made in order to find the best regression models for the willingness to adopt private AVs and public transportation systems based on AVs, respectively, which were the dependent variables. The results of these attempts are included in Table 4 and the logic of their election was as follows: First, an attempt was made including all variables (model I). Second, several variables that have been shown to be correlated with others were gradually disregarded (models II and III). Third, considering that the former models were not satisfactory, further attempts were performed using the most promising independent variables to explain the dependent one according to the literature and the authors' criterium (models IV to VII. Note that even though individually they had initially been proposed for one particular dependent variable, all of them were tested for both). Those last models were better than the more general ones. However, some coefficients were not significant. Therefore, two additional (much simpler) models were tested, in this case for the willingness to adopt private AVs (model VIII) and public transportation based on AVs (model IX), respectively, in which the independent variables were those whose coefficients had turned out to be significant in the former attempts. Indeed, those last models, based on safety perceptions and AV attraction, resulted in the most suitable ones for each of the dependent variables. The goodness-of-fit of each of the models was first assessed via the -2log-likelihood, which is a measure of the model fit. In particular, the likelihood function represents the probability of observing the gathered data given the parameters of the model. When comparing different models, the model with the lower -2log-likelihood is generally considered a better fit for the data. For both dependent variables, we have compared all models with the intercept-only model, i.e., a model that includes no variables. Additionally, all models with a significant chi-square statistic (p <unk> 0.0005) imply an improvement over the baseline intercept-only model. We have also checked the goodness-of-fit using Pearson's chi-square statistic for the model. It aims to test whether the observed data are consistent with the fitted model. Assuming (null hypothesis) that the fit is good, a large p-value, i.e., the acceptance of this hypothesis, indicates that the data and the model predictions are similar and, thus, that the model is good. Finally, we have assessed the proportion of variance in the outcome that can be explained by the independent variables through the Nagelkerke pseudo-R value. The greater this value, the better predictor the model is. Focusing on the best models, i.e., model VIII and model IX, it can be seen that an a priori appeal of AVs and the revealed safety perception while travelling in them are key factors to predict the public's willingness to shift to AVs, both private and in the form of public transportation. Additionally, trust in AVs, represented by respondents not fearing for the safety of their families traveling with them, has been shown to influence private AV adoption. In the case of AV-based public transportation solutions, the general safety perception when traveling on current public transport would also have a positive impact. Finally, the significant factors extracted during EFA were used as new stand-alone independent variables. Both driver features and driver maturity were shown to influence the willingness to shift to AVs, as Table 5 shows. Particularly people with long-standing driving experience and appropriate driving behavior seemed to be less likely to adopt them. --- Conclusions This paper aimed to contribute to the state of knowledge regarding AVs' input into more sustainable mobility. In particular, we aimed to analyze to what extent they would be useful for society. We started with the statement that the main condition for usefulness to become reality is that users accept AVs as a transportation mode. Additionally, considering other characteristics of sustainable transportation such as efficiency and respect for the environment, we wanted to find out if there were differences in the adoption of AVs depending on the form of their implementation, namely as vectors for a new public transportation system or as private vehicles. Note that a high number of private vehicles driving with low occupancy increases consumption and the likelihood of congestion. Thus, a transportation system based on individual AVs would not be completely sustainable. The main conclusions drawn from the various analyses conducted are as follows: --- • For both types of AV-based solutions, the results showed that issues related to safety were those with higher impacts on adoption. This influence can be found in previous works (e.g., [11,13]). However, safety perceptions of AVs were similar across genders in this study, while other authors found females to be more safety-sensitive (e.g., [19]). It could be hypothesized that women willing to participate in an online survey on AVs do not represent the whole female spectrum. A deeper analysis in this regard would be desirable. --- • Particularizing for the case of AV-based public transportation systems, people already feeling safe when traveling on traditional public transportation were more willing to adopt them. This finding is consistent with other studies suggesting that the users of public transport who act as passengers are more likely to accept being passengers in driverless vehicles [21]. --- • Previous attraction towards AVs also played a positive role in AV adoption, as demonstrated in previous works (e.g., [12,13]). --- • Conversely, other variables, such as the value of time (in the form of willingness to take advantage of travel time to perform any kind of task), did not show any effect on AV adoption. Previous works in the literature did find an influence of these variables (e.

Autonomous vehicles (AVs) have been proposed as vectors for more sustainable mobility, the adjective "sustainable" being understood as the union of usefulness for society (i.e., efficiency), environmental friendliness, and economic viability. The realization of social usefulness implies a primary condition, which is AVs' adoption on the part of the population. This paper is aimed at identifying factors contributing to society's willingness to shift to two different AV implementations: as private vehicles or as the basis for public transportation systems. To this end, data collected from a survey were analyzed using exploratory data analysis, exploratory factor analysis, and ordinal regression. Safety-related variables as well as the a priori appeal of AVs turned out to be the most influential for the adoption of both solutions. Therefore, informative campaigns focusing on the safety benefits of AVs could boost the modal shift. Further research starting from these hypotheses and accounting for the limitations of the sample used could help to resolve remaining doubts.

be found in previous works (e.g., [11,13]). However, safety perceptions of AVs were similar across genders in this study, while other authors found females to be more safety-sensitive (e.g., [19]). It could be hypothesized that women willing to participate in an online survey on AVs do not represent the whole female spectrum. A deeper analysis in this regard would be desirable. --- • Particularizing for the case of AV-based public transportation systems, people already feeling safe when traveling on traditional public transportation were more willing to adopt them. This finding is consistent with other studies suggesting that the users of public transport who act as passengers are more likely to accept being passengers in driverless vehicles [21]. --- • Previous attraction towards AVs also played a positive role in AV adoption, as demonstrated in previous works (e.g., [12,13]). --- • Conversely, other variables, such as the value of time (in the form of willingness to take advantage of travel time to perform any kind of task), did not show any effect on AV adoption. Previous works in the literature did find an influence of these variables (e.g., [16,18]). Thus, specific studies on these relationships are needed. • Additionally, the factor analysis suggested that actual driving features and experiences might influence AV adoption. Different results can be found in the state of the art with regard to the possible impact of these factors, which should be further analyzed. • Appeal cannot be easily influenced. However, safety perceptions can be, at least to some extent. Given the former results, stakeholders interested in AV adoption should try to clarify how AVs drive and their advantages from the safety point of view. Potential risks and their probabilities should also be explained. Information can shape perceptions, and informed individuals tend to make more objective decisions. Finally, it is important to recall some insights from the exploratory data analysis already explained. For example, most respondents did have a car and used it as their preferred mode of transportation. This fact could have influenced the results obtained. Also note that the sample size was limited and that the gender ratio was unbalanced, as was the rate of people living in or outside cities. Therefore, the results of this work should not be taken for granted, but as a starting point for future research. --- 33. How excited (attracted) are you in general with the idea of fully autonomous vehicles (SAE level 5)? • Very excited In the event of an expected crash between the automated car and pedestrians or another car, how do you think the car should react? • Trying to save the largest most amount of people • Trying to save its own passenger/s • Trying to save the pedestrians or other cars involved, no matter their number. • Trying to save the youngest people involved 45. Which autonomy level would you choose for your car, disregarding differences in the price? • Non-autonomous at all • With basic driving assistance systems (SAE 1-2) --- • Partially (conditioned) automated (SAE 3) --- Data Availability Statement: The data used to perform this research are kept by the authors, who may transfer them for academic use in particular cases and with the permission of the participants. --- Conflicts of Interest: The authors declare no conflicts of interest. --- Appendix A This appendix contains the English version of the survey. Text in quotation marks was the introduction. "The introduction of highly automated vehicles (AVs) will represent a milestone in the evolution of transportation and personal mobility. AVs are expected to significantly reduce accidents and congestion, while being economically and environmentally beneficial. However, much work remains to be done to make these improvements a reality. Furthermore, it remains to be defined which scenario is the most positive in terms of the level of automation truly necessary, the modalities of cooperation between autonomous vehicles, which type of vehicles should be automated (e.g., public transport or private vehicles, vehicles transporting people or only goods), the most suitable driving environments (e.g., city versus motorway), etc. This survey tries to study people's willingness to use automated transport, and the reasons/conditions why they would or would not like to use it. Before starting the survey, a little piece of information is provided so that everyone is able to answer the survey properly. Afterwards, the survey will not take more than 10 min. The Society of Automotive Engineers (SAE) defines 6 levels of driving automation ranging from 0 (fully manual) to 5 (fully autonomous), as can be seen in the following figure: There is a difference between autonomous and automated vehicles. Autonomous vehicles refer to automated vehicles with a SAE 5 level of automation, which means full automation. This is usually overlooked in non-scientific media or advertisement, where the adjective "autonomous" is indiscriminately used. Note that AVs are expected to behave cooperatively and to exchange information among them, with the infrastructure and with other devices (e.g., mobile phones). The aim of this cooperation is to improve mobility in terms of efficiency, safety, emissions and consumptions." What is your gender? Prefer not to answer

Autonomous vehicles (AVs) have been proposed as vectors for more sustainable mobility, the adjective "sustainable" being understood as the union of usefulness for society (i.e., efficiency), environmental friendliness, and economic viability. The realization of social usefulness implies a primary condition, which is AVs' adoption on the part of the population. This paper is aimed at identifying factors contributing to society's willingness to shift to two different AV implementations: as private vehicles or as the basis for public transportation systems. To this end, data collected from a survey were analyzed using exploratory data analysis, exploratory factor analysis, and ordinal regression. Safety-related variables as well as the a priori appeal of AVs turned out to be the most influential for the adoption of both solutions. Therefore, informative campaigns focusing on the safety benefits of AVs could boost the modal shift. Further research starting from these hypotheses and accounting for the limitations of the sample used could help to resolve remaining doubts.

Introduction Diabetes and hypertension rank among the chronic diseases that are very difficult to control. In Africa, particularly in Cameroon, traditional healers and a portion of the rural population possess a considerable patrimony of traditional knowledge on medicinal plants, heritage of their ancestral uses. So, great percentage patients in Cameroon live below the poverty line and have many difficulties in meeting fundamental living needs that include health services. As a result, some diabetic and/or hypertensive patients stop treatment and shift to herbal remedies of unconfirmed efficiency and safety [1][2][3]. Cameroon has a very rich cultural diversity with different traditional systems of medicine that need more evidence-based studies on plants, crude extracts and/or purified phytocompounds. Natural products have been the source of most active ingredients in western medicines. Utilization of natural products as pharmacological tools could lead to the development of a number of new major therapeutically active compounds [4]. Reports on potentially new chemical structures or pharmacological effects of medicinal plant extracts are mushrooming. Many hypoglycemic plants have been identified, and several of them induce the beta cells secretion of insulin. Diabetes and hypertension treatment requires plants with new mechanisms that may be able to reduce insulin resistance, act like injectable insulin at the tissue level, prevent the destruction of beta cells by antibodies or anti-illots beta action, regenerate beta cells, prevent the expression of specific diseaserelated genes, favor the intestinal absorption of glucose, and correct the abnormalities of the renin-angiotensin II-aldosterone system [1,5]. Therefore, sampling in different tribes may help in the identification of several plants with medicinal properties. The inter communities ethnobotanical information is essential to determine plants diversity, their interspecific variation, and the selection of good useful species. Increasing the awareness of traditional healers to perceive the eventual problems and solutions related to the plant materials used to treat diabetes and hypertension in different tribes, is essential for the identification of the best antidiabetic and antihypertensive plants. In this study, we identified or analyzed the medical potential of indigenous antidiabetic and antihypertensive plant species. We also endeavored to identify numerous medicinal plants by prospecting many traditional systems of medicine in Cameroon [6]. Hence, we answered some key research questions that included: are recipes used against diabetes and/ or hypertension varied significantly between socio-cultural groups located in the same phytogeographic unit of the same phytogeoraphic region (Figure 1). In other words, do the uses of medicinal plants easily transmitted from ethnic group to another? Do the populations --- Methodology The process used to collect information on medicinal plants and treatment of diabetes and hypertension began by a field work based on harvest and identification of plants of interest in each ethnic group of interest. During this step ethnopharmacological details on preparation of recipes were collected from traditional healers. This description focused on the mode and the time of preparation, the mode of administration, the posology, the duration of treatment, the undesirable or secondary effects, the toxic effects, the diet and the disease treated. It was followed by the confirmation of botanical identification that was done at the National Herbarium of Cameroon. Voucher specimens were deposited in the Institute of Medical Research and Studies of Medicinal Plants in Yaounde, Cameroon. The research of usage similarity was conducted by comparing 2 by 2 recipes or plants collected in cultural communities living in the same phytogeographic unit. The list of plants and recipes per tribe was constituted. The regrouping of tribes by phytogeographic units was realized by superposing the Letouzey' vegetation map as described by Achoundong [7] and the location of the different socio-cultural groups of Cameroon [8]. The same methods were used for identifying the different socio-cultural groups. To answer research objective we interviewed a total of 1,131 persons belonging to 58 socio-cultural groups distributed in Cameroon territory (Figure 1). --- Importance economic evaluation of forest plants sold in Yaounde markets Investigations were carried out on medicinal plants harvested and sold by 25 people in some highly populated market areas of Yaounde. Different plant packagings were previously identified, including bags or strings for all vendors and heaps for retailers (Table 1). This information Relationship between socio-cultural groups of interviewees and different phytogeographic units of Cameroon [6,9]. permits to determine the annual average profits of each category of sellers. Determination of annual average profits of different sellers. There are three categories of sellers: the delivery men, permanentretailers and collector-retailers. The delivery men, who for most of them were native of the plants' harvest zones, and relative of permanentretailers, sold the plant materials collected that include: barks, fruits, whole plants, rhizomes, and leafy stems. Annual average profits of a delivery man: Annual average profits of delivery man = Annual average selling price of all delivery men for all plants -their annual transport and hotel fees /6 (Number of recorded delivery men) Annual average profits of a permanent-retailer. Annual average profits of a permanent-retailer = Annual average selling price of all permanent-retailers for all plants -their annual average purchase price of all plants-[their annual storage fee + their annual guarding fee + their annual tax] / 14 (Number of recorded permanent-retailers). --- Annual average profits of a collector-retailer Annual average profits of a collector-retailer = Annual average selling price of all plants for all collectors-retailers-[their annual storage fee + their annual guarding fee + their annual tax +their annual transport fee]/ 5 (Number of recorded collectors-retailers). --- Results and discussion --- Important roles of socio-cultural groups in ethnobotanical assessment Comparisons of recipes between socio-cultural groups located in the same phytogeographic unit: Eighty six point ninety-seven percent (86.97%) of recipes or plants comparisons' cases between socio-cultural groups located in the same phytogeographic show that tribes use significantly different plants or recipes (Table 2). Individuals of the same origin, the same culture and the same history live by groups more or less close. For this reason a survey conducted in several sociocultural groups can permit to identify various recipes. Soudanian altitude sector; WSSS: Woody Soudano-Sahelian savannahs; WSGS: Woody Soudano-Guinean savannhas. The individuals of the same origin, the same culture and the same history live by groups more or less close. We can conclude that a survey conducted in several socio-cultural groups can allow identifying several plants or recipes. Therefore, floral, climatic, edaphic and socio-cultural differences between regions (2 X 2 compared) are found in the use of antidiabetic and/or antihypertensive plants or recipes. This conclusion is more valuable if the list of recorded plants is representative. Research on the representativeness of plants recorded for antidiabetic and/or hypertension treatment in Cameroon. The number of interviewees' socio-cultural group may be weak, at first sight. But it is in general widely superior to a minimum doorway of four individuals recommended by category for example socio-cultural group, within the context of ethnobotanical surveys [9,10]. The 1.131 interrogated Cameroonians appear negligible when compared with the total population of Cameroon estimated to 13 600 000 of habitants [11]. In this case the rate 0.0832% of participation is very weak. For practical reasons, the minimum doorway of sample difficultly attains 1% of entire population [10]. Many of reticence and distrust from the informants make this minimum doorway more reduced in the area of medicinal plants. That enables us to address the following question: after all, is the list of plants recorded representative of the set of plants used in Cameroon against these two diseases? The answer of this question stands on the study of the growth of the number of recorded plants in relation to the interviewees of each phytogeographic region (Figure 1) Research on the representativeness of plants recorded for antidiabetic and/or hypertension treatment in Cameroon that include Coastal humid forests, Continental humid forests and Guinean and Soudano-zambazian savannahs, in consideration of the fact that the regions compared 2 by 2, show a significant difference. We consider that the interviewees were drawn by lot (without drawback): 14 groups of 20 andone group of 13 in the phytogeographic region 1; 13 groups of 20 andone group of 17 in the phytogeographic region 2; 28 groups of 20 andoneinterviewer in the phytogeographic region 3. The curbs in each case are illustrated in the Figure 2. The best adjustment of the trend of each of these curves is a logarithmic function with the following equation: The analysis of these curves reveals that the number of recorded plants increases with that of the interviewees until a certain level of interviewees (240 -293) in region 1, (200 -277) in region 2 and (480 -561) in region 3. Beyond these levels, any increase of the number of interviewees doesn't lengthen the list of antidiabetic and/ or antihypertensive plants. The list of recorded plants is therefore representative of the set of plants used in Cameroon against the two diseases. We must remark that this is true if the new specimen is distributed in the way that the sample repeated present the same characteristics like the first one, in ethnical point of view; in admitting that each socio-cultural group has a delimited territory, and the same affinity with others that is not taken in to consideration. and antihypertensive, seventy-four (74) are used against hypertension, and fifty-nine (59) are antidiabetic. These recipes were derived from 203 plants belonging to 68 families [9]. Among the 203 recorded plants, 33 have been shown to be very interesting because of their clinical outcomes in 182 patients who used them in self-medication [5]. Ecological importance: Wide ecological plasticity plants that include Momordica charantia, Voacanga africana, Canarium schweinfurthii, Ceiba pentandra, Rauvolfia vomitoria, Allophyllus africanus and Kigelia africana currently used in most socio-cultural groups. Food, cultivated and local markets sold plants: Fifteen recipes are derived from native or introduced plants, which are cultivated for commercial purposes. Native species used also for foods include: Solamum melongena, Corchorus olitorius, Cucumis metulliferus and Vigna unguiculata, for other medical uses include: Vernonia glabra, Aloe barteri, Aloe buettneri and Aloe schweinfurthii. Introduced plants are Persea americana, Cymbopogon citratus, Catharanthus roseus, Brassica oleracea, Allium cepa, Allium sativum and Citrus grandis. Wild species which stem barks are sold in the Yaounde markets include: Mammea africana, Ceiba pentandra, Cylicodiscus gabunensis, Guibourtia tessmannii, Spathodea campanulata, Ricinodendron heudelotii, Zanthoxylum zanthoxyloides, Entandrophragma spp, Newbouldia laevis, Morinda lucida, Bridelia micrantha, Nauclea diderrichii, Antrocaryon klaineanum, Pterocarpus spp, Garcinia lucida, Fillaeopsis discophora, Hallea stipulosa and Pteleopsis hylodendron. Local markets in urban areas nowadays are refuges for traditional knowledge and places for disseminating of new knowledge and practices about medicinal plants [12]. Like in India, wild edible plants are widely consumed in the daily diet of local people in Cameroon [1]. They are critical for the sustenance of ethnic communities and also as source of income [13,14]. --- Socio-economic importance: Commercialization of antidiabetic and/or antihypertensive plants and possibilities of sustainable development of many tribes. --- Calculation of storage fees, hotel fees and tax by categories of sellers: Each of the permanent-retailers or collector-retailers pays 200 FCFA for municipal tax per day and 100 FCFA for guarding fee per day, in addition to 40 000 FCFA for annual tax. The year has 360 days for permanent-retailers and 300 days for collector-retailers for which 2 months are used for supplies. The collector-retailers dwell in the collection sites. The 14 permanent-retailers spend per a year for all the fees: ((200 FCFA + 100 FCFA) x 360 + 40 000 FCFA) x 14 = 2,072,000 FCFA. These expenses are added to 4,918,300 FCFA that is their annual total purchase price of all sold plants. A back is the unit for selling medicinal plants by the delivery man and a heap is the unit of selling for retailers. The value of each unit depends on the importance of the resource and varies from 3,000 FCFA to 10,000 FCFA for a back and from 300 FCFA to 1,000 FCFA for a heap. The 6 delivery men spend per year 588,895 FCFA for transport and hotel fees. The 5 collectors-retailers spend per year for all the fees: ((200 FCFA + 100 FCFA) x 300 + 40,000 FCFA) x 5 = 650,000 FCFA. They also pay 209.335 FCFA for transport fees. Therefore, their total annual expenses for all plants is 650,000 FCFA + 209,335 FCFA = 859,335 FCFA Table 3: A seller annual average profits on all antidiabetic and/or anti-hypertensive plants sold in Yaounde markets TAASPP: Total annual average selling price per category of sellers; TAAPPSC: Total annual average purchase price per category of sellers; TAESC: Total annual expenses per category of sellers; TAAPSC: Total annual average profits per category of sellers; AAPSC: Annual average profit of a seller per category. This Table 3 shows that a collector -retailer earns more than two times than a permanent-retailer and 148,465,500 FCFA more than delivery man. Like the delivery man the collector-retailer doesn't buy the resources. We realize that the natural harvest and the retail sell are advantageous for the local commercialization of medicinal plants. Therefore, for the same resources sold in the market, personal collection following by retail may be highly recommended. Meanwhile, the formula tax, storage fee and municipal fee significantly decrease the income of permanent-retailers. Other profits are made from the sale of non-antidiabetic and/or anti-hypertensive medicinal plants. In general sellers of medicinal plants can fulfill their financial needs. However, plant harvesting pose a threat to the environment as it becomes more intense in areas far from urban and rural agglomerations. In Yaounde, prices vary according to season and the distance of the zones of harvest. In Mokolo market the resources are from Lekie Division (Okola, Evodoula and Monatele Subdivisions) and sometimes from Nyong-and Ekele Divisions. In Rond-point Longkak market, plants are from Mbam and Inoubou and Mbam and Kim Divisions. The Mvog-Mbi market is supplying by people of Mefou-Afamba, Mefou-Akono, Ocean and Nyong and So'o Divisions. --- Comparison of recorded plants in the three phytogeographic regions: The statistical analysis has shown that the list of recorded plants is representative of the set of potential antidiabetic and/or antihypertensive species of Cameroon (Figure 2). Also, a significant difference in plants was observed between the phytogeographic regions compared 2 by 2. Only thirteen plants that includes: Rauvolfia vomitoria, Momordica charantia, Allium cepa, Cissus quadrangularis, Laportea ovalifolia; Garcinia afzelii, Bidens pilosa, Perseaamericana,Catharanthus roseus, Cymbopogon citratus, Margaritaria discoidea, Canarium schweinfurthii and Voacanga africana were common for the three regions. --- Ethnopharmacological preparation and ethnomedical administration of the standard recipes: Almost all the plants cited in this work were prepared and administrated as follow: 100 g of leafy stems of herbal species or 200 g of stem barks were boiled in 4 liters of water, for 25 minutes, each patient was then asked to drink a glass of extract (250 ml) in the morning; mi-day and evening, for 10 days. Specific preparation of each of these plants is detailed in [3,6]. No secondary and undesirable effects were recorded during the treatment. But the prolonged use of Rauvolfia vomitoria and Cissus quadrangularis can induce vomiting. Maceration of the same quantity of harvested material in the same volume of water, for a whole day, is also allowed for the following plants: Allium spp (bulb), Brassica oleracea and Aloe spp. In these cases, the same posology is respected [9]. --- Conclusion This ethnobotanical study was designed to collect relevant information on medicinal plants or recipes used by various sociocultural population groups living in different phytogeographic units of Cameroon. From a two by two comparison of social-cultural groups located in the same same phytogeographic unit, it was found that a significant difference exists in 86.97% of medicinal plants or recipes comparisons' cases. A total of 208 recipes were identified, including 75 that were used for diabetes and hypertension treatment, 74 for hypertension alone, and 59 for diabetes alone. Also, 203 medicinal plants were identified including 33 were cultivated and marketed by 25 farming families who also sold the barks of 18 wild plants. Their engagement in sustainable agriculture and marketing of medicinal plants is helping to improve their socio-economic status. However, many of the antidiabetic and antihypertensive plants are seldom collected and cultivated, although comprehensive assessment has shown that many of them have high clinical outcomes. Hence, further research should be conducted to extract, purify and characterize the active ingredients of these plants. Despite the representativeness of the list of recorded plants identified for diabetes and hypertension treatment in Cameroon, there is a need to improve the methodology for identifying new medicinal plants.

Diabetes and hypertension rank among human diseases that are very difficult to control. The medicinal material of Cameroon can provide much information on ethnic folklore practices and traditional aspects of therapeutically important natural products. Cameroon has a very rich cultural diversity with different traditional systems of medicine that need more evidence-based studies on both crude extracts and purified phytomolecules. Therefore, an ethnobotanical study was conducted on 58 socio-cultural population groups living in different phytogeographic units of Cameroon in order to collect various medicinal plants or recipes. A two by two comparison of social-cultural groups of the same phytogeographic unit indicated a significant difference in 86.97% of medicinal plants or recipes comparisons' cases. A total of two hundred and eight recipes were identified, among which 75 were used for diabetes and hypertension treatment, 74 for hypertension alone, and 59 for diabetes alone. Also, two hundred and three plants were identified among which 33 were cultivated and marketed by 25 farming families engaged in integrated agriculture and selling of antidiabetic and antihypertensive plants to enhance their socio-economic status. Tsabang N (2017) Socio-cultural contribution to medicinal plants assessment and sustainable development: case of antidiabetic and antihypertensive plants in Cameroon

Introduction Current psychiatric research assumes the mechanisms of mental disorders can be understood in terms of neurobiology, especially brain circuitry. However, mental health problems clearly depend on developmental processes of learning and adaptation through ongoing interactions with the environment. Human environmental niches are socially and culturally constructed. Symptoms or disorders emerge in specific social contexts and predicaments that cannot be fully characterized in terms of brain function but require a larger ecological systems view. Causal processes can begin anywhere in this larger ecosocial system. In particular, individuals' narrative self-construals, culturally mediated interpretations of symptoms and coping strategies, as well as the responses of others in the social world, can play a crucial role in the mechanisms of mental disorders, illness experience, treatment response, and recovery. In this paper, we outline the conceptual basis and practical implications of this hierarchical systems view for psychiatric theory and practice. We argue for the importance of adopting a cultural-ecosocial systems view that understands the brain as situated in the social world and as part of larger, self-reflexive systems that are embodied and enacted through language and other cultural practices (1). This view builds on work in systems biology, social epidemiology, developmental psychology, anthropology and 4E cognitive science to provide a process-based view of the dynamic interactions of subjective experience and social context. We use the term 'ecological' here in a way that is related directly to ecology [as the study of biological organisms in relationship to their physical environments (2); see (3,4)] but with the recognition that for humans, the environments in which we are embedded are fundamentally social and cultural. What flows through these organism-environment systems is not just energy or material (as is the case in typical ecological analysis) but also information, which is essentially relational (5). The material and symbolic (informational) dimensions of our environment are closely related. We inhabit socially constructed niches that enable communication and cooperation (6). We employ cultural knowledge and practices to navigate these niches, which are both local and extended through time and space. In the process, we both actively reconfigure these niches (7) and are reshaped by them at neurobiological, cognitive and social levels (8). The cultural-ecosocial view encourages us to consider how cognition and experience depend on the dynamics of the system comprising organism and environment. By emphasizing systemic processes, this view takes a step away from narrow concepts of mechanism that assume the total decomposability of a system into its parts (reductionism), with the recognition that the ways that the parts or constituents of systems are spatiotemporally arranged and connected give rise to new dynamics. System dynamics arise from connectivity, organization and interactions not simply from the properties of the components. Interactions between components may change the properties and function of each of the components as well as the dynamics of local and larger networks. In the sections that follow, we first provide a brief genealogy of systems thinking in psychiatry and outline the specific contributions of the existing frameworks that we draw from. We then discuss the nature of hierarchical organization in biology before turning to a discussion of multilevel explanation in psychiatry. The next section argues that 4E cognitive science can provide a path to multilevel integration through a cultural-ecosocial systems view. We then illustrate with a case vignette how this approach can be applied to integrative clinical case formulation. The cultural-ecosocial systems approach includes patient's experience, self-understanding and agency, as well as social structural processes, in explanations of symptoms, disorders and distress. Finally, we discuss the implications of our approach for psychiatric practice. We are calling for a change in psychiatric theory, research and practice that resists the reification and over-simplification of mental health problems in terms of discrete diagnostic entities by focusing on system dynamics that include individuals' experience and meaning-making as well as the socialcultural contexts in which the person is embedded and from which psychiatric disorders emerge. --- Systems thinking in psychiatry The concept of system is associated with Enlightenment views of knowledge and has been a central trope in modernity associated with ideas of order and control (9). However, a more abstract notion of system has served as of way to identify important analogies and formal correspondances among diverse phenomena. In this usage, a system is a structured ensemble of parts or processes (components, actors or agents) that interact in ways that allow the structure to persist over time and exhibit distinctive behavior or dynamics. The recognition that very different kinds of ensembles may display similar dynamics reflecting their organization led to the development of general systems theory (10)(11)(12) and cybernetics (13). The application of systems theory received new impetus with the development of computational approaches to modelling dynamics. Mathematical analyses and computational modelling revealed complex dynamics emerging from even simple systems spawning the development of subfields of nonlinear systems theory, and popular re-branding under the banners of "chaos" and "complexity theory" (14)(15)(16)(17)(18)(19). The focus on dynamics supports an ontology in which systems are characterized not in terms of their constituent parts and structures but rather in terms of interactional processes (20,21). The concepts introduced in general systems theory and cybernetics were applied widely to modelling behavioral, biological, ecological and social-economic systems [for overviews see (10,22,23)]. Efforts to understand biological processes at genomic and cellular levels led to the development of systems biology (24). In this view, the function of components of biological systems like genes, organelles, cells, tissues and organs can only be properly understood by considering their relation to the dynamics of the larger system as a whole. Understanding these dynamics holds great promise for improving clinical approaches to the assessment and treatment of myriad complex medical conditions (25,26). Systems thinking has a long history in psychiatry, going back to the development of the notion of homeostatic regulation of physiological systems in the work of Walter Cannon (27), and some of the early applications of systems theory (28). Psychiatrists and neuroscientists were key figures in the development of cybernetics in the 1940s and 50s (29). This work aimed to model learning and adaptation in mechanistic terms and identify forms of pathology with specific types of dysregulation of adaptive systems. Subsequent work applying systems thinking to understand psychopathology was inspired by systems biology (32), the cybernetics of behavioral control systems (33)(34)(35)(36), complexity theory (37,38), and recognition of the impact of social-structural determinants of health (39). A recent version of control systems modelling can be found in the active inference approach to explaining specific forms of psychopathology (40). While focused initially on neural processing, active inference can be readily extended to consider interactions with the environment and social networks, (41)(42)(43)(44). Systems theory and cybernetics were central to the development of family therapy (45)(46)(47)(48). Families were viewed as self-regulating systems comprised of individuals in interaction with each other (49). These interactions are influenced by individuals' characteristics but also reflect spatial, material and symbolic structures as well as the social practices, norms, rules, and rituals that constitute family life. The family system is constituted both by the individuals who are its members and the community, society or culture that configures and constrains its structure and identity. The family system thus serves the needs of its members and of the larger society in which it is embedded-and these diverse needs may sometimes conflict with each other. While there have been substantial efforts to elaborate sets of dimensions, typologies, and measures to characterize the structure and dynamics of families [e.g., (50)(51)(52)(53)(54)], to date, none have achieved wide acceptance or clinical application. The interactional view of family systems has been extended to consider larger social networks and structures (55-58) but this is also an unfinished project (59). In the 1970s, second-order cybernetics theory emphasized the role of recursivity, self-reference and self-construction (autopoiesis) in living systems (60). This opened the way to a deeper engagement with social, cultural and linguistic processes of meaning making (39,(61)(62)(63). Agency, subjectivity, and narrativity are given central place in systems approaches that acknowledge the role of communication, storytelling and self-reflection in human experience (45). There are many interconnections among these different approaches to systemic thinking in psychiatry. This reflects both their shared genealogy<unk> with common ancestors, and cross-fertilization among disparate strands. These lines of work are complementary and each can address some of the limitations of the others. In particular, systems neurobiology focuses on multilevel processes in the nervous system but does not sufficiently consider the social environment or treats it simply as a modulator of neural processes rather than as constitutive of brain function. Systems biology has been extended to consider biosocial interactions, but this work often does not specify the psychological processes of meaning and experience that mediate bodily and social interactions (64). Krieger's (65,66) ecosocial epidemiology uses the construct of embodiment to emphasize the biological effects of structural inequities (i.e., how adversity gets "under the skin") but also does not clarify psychological processes. Bronfenbrenner's (67)(68), bioecological psychology emphasizes the dependence of developmental processes on environmental context but its application has not always considered the interaction of the multiple levels of social organization in which individuals and families are embedded (69). Ecocultural approaches grounded in ethnographic methods have provided ways to characterize the culturally constructed meanings and practices that constitute lifeworlds and developmental pathways (70). The notion of an ecology of mind, introduced by Bateson (61, 71) views cognition as emergent in loops of individuals interacting with the environment and through interpersonal communication with other humans in a social system (72). The many strands of 4E cognitive science develop this perspective in terms of processes of embodiment and enactment that involve social embedding and extension in the world (73)(74)(75)(76). The cultural psychiatric perspective emphasizes the interactions of individual and collective meaning making and the social-political contexts of institutional power and practice that create cultural niches and affordances (77). Computational methods allow us to put aspects of each of these approaches together in an overarching model that can reveal system dynamics (78). The novel aspects of our approach that distinguish our framework from previous work include: the explicit integration of culture (as embodied background knowledge and enacted situated practice); the characterization of basic psychological processes of subjectivity, narrativity, and agency in terms of embodiment and enactment; and a focus on the dynamics of multilevel biological, cognitive and sociocultural looping effects as potential mechanisms of pathology and targets for intervention. --- Hierarchical systems theory in biology Biological systems are hierarchically organized, with components that are arranged in ways that give rise to stable structures with new properties and processes (20). For example, the metabolic processes of the cell depend on the spatial organization of enzymes on its membranes. The computational functions of the brain depend on its hierarchical structure of networks and nodes (79). This organizational process is recursive and new control processes emerge as a result of the hierarchy (34). This hierarchy includes the social environment which emerges as part of specific arrangements of relationships with others through social norms, rituals, institutions, and practices-and which, in turn, shapes the development and functioning of the individual. The notion of hierarchy sometimes conjures images of domination or oppression. However, as we use it here, hierarchy does not involve value judgments about degree of importance, power or privilege but refers to specific forms of organization of systems (80). Some philosophers are critical of the idea of hierarchy and levels in living systems because they see this as imposing a misleading model or metaphor on phenomena that are fluid, shifting, or 'holistic' (81). 1 Others are concerned about the notion of 'top-down' causation, arguing that causal processes can only involve same-level processes that are materially linked (83). In reality, top-down causation is common in complex systems and is central to organismic biology (84). There are many types of organization that can be discerned in the world or applied to experience and the utility of concepts of hierarchy and levels depends on the specific question, problem, object of interest and pragmatic task at hand (85,86). 2 The notion of hierarchy is used in multiple ways in biology that include subsumption, composition, scale, causality, and control (88,89). Hierarchy implies organization into levels, but the significance of these levels differs in each of these versions of hierarchy. In biology, each of these notions of hierarchy is useful but the one that is most important for an ecosocial systems view in psychiatry is that of control hierarchies. Subsumption hierarchies are classifications in which something is seen as a member of instance of a larger category. An example is a Linnean taxonomy of species taxa. The logical relationship between levels can be captured by set theory. The elements of progressive levels are sets of the prior level's sets. Elements at lower level may be viewed as concrete instances, while higher levels are abstractions, or each level may have a kind of ontological identity (90). A lower-level instance can stand metonymically for the whole. But the way that elements are 1 There is no doubt that notions of hierarchy reflect sociomoral and political values (73), but this does not vitiate their use as technical concepts in science or other domains. 2 Of course, there is legitimate concern about the need to recognize oppressive structures in society but these are not simply due to hierarchically structured systems. Oppression can arise not only from the dominance of one group over others but from systemic processes that occur at multiple levels in the system. For example, collective norms and public discourse can legitimate discrimination, micro-aggressions and social exclusion with significant effects on the health of minorities. Hence, it is important to distinguish between hierarchical structure as an organizational feature of complex, self-organizing, goal-oriented systems and oppressive processes that make use of hierarchies, but also other aspects of social structure and everyday practice, to exert power in ways that create inequity. Ecosystems and social systems might be better characterized in terms of the concept of "panarchy" in the sense that they exhibit both top-down and bottom-up causation often on different spatial and temporal scales (78). The potential for conflict between these levels of causation and control leads naturally to a consideration of the dynamics of social power. Scale refers to the number or the size of the assembly relative to its components, spatial or temporal span. Scale differences can be continuous or discrete. Some biological and social network-based phenomena are'scale free'; that is, the same structural organization and dynamics are observed at multiple scales or else scale-up in a quantitatively predictable way (91,92,93). This allows dynamical system models to be applied in an iterative way to characterize processes across these networks at multiple scales. However, many physical and biological phenomena are not scale free; that is, size matters (94)(95)(96). The sheer number of elements, their topological arrangement or connectivity, and their spatial or temporal extent can give rise to new dynamics. In this case, the emergence of new dynamics marks a new level in a hierarchical structure. --- Gómez In compositional hierarchies, the focus is on part-whole relationships (97). The parts are building blocks that are arranged in spatiotemporal structures that create a new level of organization. Bricks are laid to build a wall; walls joined to build a room; rooms are concatenated to build a house; houses are arranged to create a neighborhood. The process of composition may involve different kinds of arrangement at each level and similar processes may be involved in stabilizing the structures (e.g., mortar may be used to build walls, to join them into rooms, and to join rooms into a house). However, different processes (reflecting other properties of the components or additional components) may stabilize structures at different levels (buildings might be joined by mortar to build a wall and walls might be joined by interleaving bricks at a corner, or by at angle brackets made of metal). Depending on our focus of study, the level and processes we need to explain a phenomenon will shift. Thus, if we are looking at the stability of a house we will be interested in the strength of bricks and mortar bonds, and the buckling properties of columns and frames; whereas, if we are interested in neighborhood stability, we will need to consider parameters at other compositional levels like street layout, greenspace, and social relationships among inhabitants. However, we may find that house stability and neighborhood stability significantly affect each other because of mechanisms that link these through social and economic processes such as house pricing, gentrification, neighborhood pride and upkeep. Compositional hierarchical organization is central to biology and essential to phylogeny, ontogeny, and adaptation to new environments because biological systems build on existing structures by preserving, re-organizing, and re-purposing components (98). In biology, there are multiple compositional hierarchies, but the main line follows from the ways that processes are stabilized to create a hierarchy of material structures (99): molecules are joined to make macromolecules (through chemical bonds); macromolecules are arranged in space (with the aid of membranes and other macromolecules) to produce organelles; organelles are arranged in space (again with the aid of membranes, macromolecules and other organelles) to create cells which have metabolic cycles; cells are organized into tissues which have biomechanical and other functional properties; tissues are organized in organs which can perform multiple functions related to their structure and anatomical location; organs form physiological systems, which have properties related to interactions between the organs they connect; physiological systems constitute organisms; organisms form communities; and diverse communities in environmental context constitute ecosystems. Causal hierarchies reflect arrangements determined by mechanisms or processes that produce a given effect (100,101). The directionality of the link (or irreversibility of the process) establishes an ordering. The ordering of causes leading to outcomes which are causes of subsequent outcomes provides a sequential structure that can be described as a chain of cause and effect. Of course, most processes have multiple causal contributors that interact and result in different partial orderings or lattice structures that may have a layered or hierarchical structure. Moreover, multiple causes may independently lead to the same outcome (equifinality), and single causes may lead to multiple outcomes (multifinality), presumably reflecting the influence of other historical or concurrent causal factors. Finally, the assumption of unidirectionality at one causal level may not hold when the larger system of relationships is considered. Most biological systems involve mutual or circular causality or feedback loops. Indeed, circularity (autocatalysis, self-assembly or autopoiesis) is essential to what characterizes a system as living (102)(103)(104)(105)(106). Through such circularity and self-reference, biological systems then instantiate another form of hierarchy that involves self-regulation or control. Control hierarchies are defined in terms of successive levels of regulatory loops (107). The control systems perspective is especially relevant to understanding biological processes (and psychopathology) because it leads to a useful way of understanding function and dysfunction in terms of the goal-oriented nature of behavior and adaptation. A basic building block is a feedback loop in which a state of the organism or environment is compared with an expected (or desired) state [what Miller et al. (34) called a 'Test-Operate-Test-Exit' or TOTE unit]; the discrepancy then drives a compensatory action (either revising the expectation or acting on the world to make it better conform to the expectation). Successive levels are loops of loops. These loops can involve different processes that are best characterized as regulating information (or 'free energy') rather than energy per se (108). This is the kind of hierarchy of greatest interest in making sense of the dynamics of living systems. For living systems, these loops are characterized by a fundamental regulatory goal of maintaining organism integrity and persistence in the service of reproduction and other goals. The resultant teleodynamics distinguish living systems from other regulatory systems that lack the capacity to generate organism-specific goals and norms and to function in ways that are explicitly informed by future possibility (109,110). 3 In humans, this process extends to the self-reflexive, imaginative and cooperative processes of agency enabled by language and culture (112,113). The general idea of hierarchy then does not imply unidirectional (top-down or bottom-up) causation, linear dynamics, or reductionism. In fact, evidence for hierarchical organization is seen in many emergent phenomena. Emergence involves the appearance of new levels of organizational structure without implying loss of underlying structures or component levels (114). These new levels of organization have their own dynamic processes. The emergence of new structures with distinctive properties and of processes with new dynamics warrants the use of the concept of levels of organization and corresponding levels of description. In hierarchical systems, the function of each level can be explained not only through the interactions of its components but in terms of its relationship with both higher and lower levels. For example, the genome is a set of structures used by the cell to regulate its activity and replicate itself; the genome itself is a dynamic system that is regulated by a network of macromolecules (115). Similarly, the cells of a healthy multicellular organism serve the priorities and plans of the whole organism-sometimes to the detriment of their individual survival. 4The functions of any level in a biological system then only make sense in relation to the dynamics of the larger system, including the regulatory processes organized at higher levels. The principle of biological relativity, developed by physiologist Denis Noble (117)(118)(119), argues that in biological systems causal chains can begin anywhere within the system or hierarchy. This approach to systems biology recognizes the organizational value of hierarchy but is explicitly antireductionist in the sense that both lower and higher levels of organization have causal efficacy and contribute to the dynamics of the system as a whole or the subsystems that constitute brains, persons, families and communities. --- Multilevel explanation In psychiatry The biopsychosocial (BPS) approach championed by Engel (120,121) promised a conceptual framework to integrate multiple levels of analysis in psychiatry based on general systems theory (12). The motivation for this was a concern to give a place in clinical theory and practice to the intrapsychic processes characterized by psychodynamic theory and patients' own experience and understanding of their condition (122). But the definition and operationalization of these level and their exact interplay in cross-level formulations, were left undetermined. Critics of the BPS, like Ghaemi (123) have argued that the framework is little more than a placeholder with no real content to guide diagnostic assessment, formulation and treatment (124). To a large extent, this claim says more about critics' failure to engage the burgeoning literatures of systems biology, psychophysiology, family systems theory, social epidemiology, and other social sciences, which can put ample flesh on the bones of the BPS model, than about any inherent limitations of a multilevel systems approach to health and illness (125, 126. The lack of engagement with this literature is evident in Ghaemi's alternative proposal that psychiatry employ mechanistic biological accounts of disorder complemented by phenomenology and a humanistic concern for patients' experience. In this approach, the causal mechanisms of psychopathology are divorced from the social world. Subjectivity and social context are acknowledged as important to ensure a humane engagement with the patient but are not seen as primary mechanisms of pathology and are taken for granted as aspects of the patient's clinical presentation that can be adequately accessed and assessed with empathy and common sense. While Ghaemi's concern that the BPS leads to "undisciplined eclecticism" seems to us to be unfounded, more valid concerns are that in practice the BPS remains mainly descriptive rather than dynamic, simply enumerating potential risk, causal or maintaining factors, without detailing causal mechanisms that could guide intervention. Perhaps this is why, despite its widespread acceptance, the BPS has failed to prevent or reverse the adoption of reductive biological explanations in psychiatry. Moreover, while the BPS was motivated by concerns to include patients' lived experience, even mental health practitioners who claim to use a BPS approach tend to neglect subjectivity and social-cultural context. This failure may reflect the lack of interdisciplinary training (127,128), the difficulties of conceptual integration (81), and the persistence of dualistic thinking (129). We start from a different premise, supported by a wealth of research in psychosomatics and sociosomatics, that insists that symptoms and syndromes in psychiatry arise from the interaction of psychophysiological, cognitive-affective, and sociocultural processes (77). Psychiatric disorders are complex, multidimensional constructs, and symptoms are more than just indices of an underlying neurobiological mechanism that can be captured by biomarkers (130,131). Psychiatric disorders emerge within loops that involve the biology of human adaptation as well as cultural practices of diagnostic labelling, health care systems and larger discursive formations. Illness experience therefore does not follow directly from pathobiology but is embedded in cognitive and social processes that mediate and modulate the translation of physiological or psychological disturbance into symptoms and behaviors. This transduction and translation occurs at multiple levels that involve symptom schemas and their interaction, interpersonal responses, narrative conventions, social positioning, the health care system, economic constraints and sociopolitical processes (132) 5. This perspective is consistent with recent work in symptom network theory, which suggests that psychiatric disorders result from the dynamic interaction of multiple symptoms each of which may have its own pathophysiology or psychopathology (133), (136). Instead of assuming that a single latent construct can explain the symptom patterns that characterize psychiatric disorders, network analysis views disorders as systems of causally connected symptoms (137). These causal connections can involve physiology, behavior, experience and interpersonal interaction, as well as the responses of social institutions and the environment. While some authors consider a network as an inherently non-hierarchical structure, causal or control hierarchies may be part of the mechanisms that constitute and connect symptom networks, not as a matter of composition (or latent constructs) but as part of causal chains or loops. The ecosocial systems view we outline in this paper extends the idea of symptom networks to include social-cultural contexts, self-reflection and narration as active causal processes (1). These multiple levels of process reflect structures that are organized hierarchically in the sense that higher organizational levels involve arrangements of structures at lower levels that give rise to new processes that require new conceptual vocabularies to describe. For example, the brain is composed of functional circuits, which are made up of neurons; the social world is made up of roles, niches and institutions which are constituted by patterned relationships among individuals, whose behavior is regulated by cognitive maps, models and affordances, social positionality, norms, and conventions (138). Each level enables processes that contribute to the causal mechanisms that underlie a particular symptom, syndrome or affliction (139). Experience, behavior, narrative self-understanding, and social interactions can all contribute causally to the dynamics of psychiatric symptoms and disorders (140)(141)(142). Even brain-based explanations of mental disorders require an appeal to multilevel systems dynamics (143). Changes in synaptic function or neural circuitry alter information processing, which in turn gives rise to changes in social behavior and experience (144). The process is bidirectional. Psychotherapy and other psychological interventions have effects on the brain (145). Changes in social behavior alter brain function in ways that may be self-sustaining or create knock-on problems in other brain systems or behavioral functions. Social environments and models of the self in context influence neurobiology, immunology and inflammatory processes (142). Beyond neurobiology, mental disorders also involve cognitive, affective and attentional processes that emerge from particular learning histories and narrative modes of recollection and selfnarration, as well as interpersonal interactions with others in one's family, community and wider social networks. These social interactions have their own dynamics that may aggravate or mitigate symptoms or create predicaments that present their own challenges to health and well-being. Social interactions can also feed back into cognitive and bodily processes in ways that amplify or diminish symptoms and distress. These loops correspond to relationships between different aspects of the organism or between the organism and the environment. Loops may result in cycles of positive and negative feedback, with effects locally as well as across the organizational hierarchy. Depending on their structure, parameters and initial conditions, loops can result in nonlinear dynamics, for example, growing exponentially, showing discontinuities, bifurcations, or other complex dynamics (146,147). To the extent that these loops have their own dynamics, they can be viewed as specific mechanisms that need to be considered in diagnostic assessment and case formulations and that can be the target of clinical intervention. Moreover, because human adaptive systems involve regulatory or allostatic processes with specific goals or set points, they may exhibit equifinality, in which, despite variations in initial conditions and ongoing perturbations, they tend to follow a predictable trajectory. Identifying these stable patterns or trajectories could provide a basis for a typology of disorders organized in terms of regulatory processes that exhibit stable attractors, limit cycles, and final common pathways. 6 If these can be identified and empirically validated, they could be used as a basis for diagnoses that are prognostic (predicting outcomes) or that indicate potentially effective treatments, and that point to specific targets for intervention. This systems-based nosology, however, will generally be quite different than simply identifying single mechanisms, causal factors, or etiologies for disorders because it involves dynamic properties of systems with looping effects. A typology of looping effects (vicious or virtuous) could complement current diagnostic nosology (149). This enlargement of frameworks would not completely supplant current nosology, which has its uses, insofar as it captures salient aspects of illness experience and can be related to prognosis or differential therapeutics. Clinical assessment routinely goes beyond diagnosis to include a problem list-some categories of which are included in the ICD and DSM-5 Z-Codes (150, 151)-and case formulation that may note contextual factors, but this process is unsystematic. Efforts to systematize the inclusion of social context and determinants of health in assessment are urgently needed. This needs to go beyond a laundry-list of factors to include dynamics. Person-centered diagnostic assessment includes characterizing strengths and resources, risk and protective factors, and relevant developmental, ecological and meaning-centred contexts (152). Attention to looping effects could be incorporated into current practice through case formulation and systemic intervention without waiting for the development of a systematic nosology. Table 1 lists some of these potential loops both within levels or domains and across levels using depression as an example. Although loops are difficulty to study, they are composed of causal arcs that can be characterized with existing methodologies. Table 1 lists many such causal arcs that linked together would result in 'loopy' dynamics. This kind of model is central to cognitive theories of depression and anxiety (179), which have led to effective treatment interventions and can readily incorporate cultural-contextual factors (180). There have been some notable successes in identifying predictors of dynamics in couple interactions (181). New experimental methods have been developed to study the dynamics of dyadic, family, and group interactions (182)(183)(184). Symptom network theory and computational modelling provide new approaches to examining looping dynamics, testing the relative strength of specific linkages and the sensitivity of network dynamics to changes in parameters that can be matched with measurable variables in research and clinical applications [e.g., (185)(186)(187)]. In clinical settings, nonlinear dynamics are commonly observed and putative explanations in terms of loops could be tested by interventions that target specific parameters (188, 189). Identifying the feedback loops that may contribute to psychopathology is difficult. Statistical methods can be used to show timelagged autocorrelations and cross-correlations in observational data that suggest feedback dynamics (190,191). Experimental methods that manipulate particular parameters or control the nature of physiological, perceptual or interpersonal feedback can provide firmer evidence for feedback mechanisms (192). Computational models can be constructed that capture some of the interactions and identify parameters that affect 6 For definitions of these terms and others used to characterize system dynamics, see: Mainzer (21); for examples of how they may be related to specific types of pathology; see: Durstewitz et al. (132). Gómez-Carrillo and Kirmayer dynamics (193). However, in practice, these usually are simplified 'toy' models that do not include many of the loops and variables present in real-world contexts. This may lead to mistaken predictions or overgeneralization. There is a need for an extensive research program of modelling built on large datasets that include potentially important individual and contextual variables (194). Applying computational models in clinical settings poses additional challenges related to the constraints of clinical epistemology. The data available for an individual patient may be very limited and not include a time-span necessary to reveal dynamics. The interventions that clinicians make are not really single-subject experiments because they occur within a context of expectations and demands that heavily constrain patients' response. The patient's own interpretations and self-construals affect the impact of any intervention and any subsequent interaction with the clinician. Hence, we need a circular hermeneutics to complement our models of circular causality (195). The system of patient and clinician must be included in the model and situated within the larger ecology of health care and adaptation in social context. Crucially, the loops relevant to clinical concerns include modes of self-construal based on cognitive, social and cultural models, institutions and practices (77). For example, the interpretation of experiences of pain, fatigue or lack of interest as symptoms of depression is a culturally shaped attributional process that leads to particular modes of coping and help-seeking (149). These attributions may be re-negotiated in clinical and other social contexts with others who may validate or contest the views of patient or physician (196). To the extent these social and clinical responses validate the individual's self-construal, they constitute a loop in which the available categories for symptom interpretation and clinical practices reinforce (198) termed "biolooping") or they may primarily involve cognitive and social-rhetorical processes that reconfigure the sense of self (173,199). Loops also may be irreducibly social or political, changing the larger environment and available narratives in which social position and structural adversity determine the causes and course of symptoms. Psychiatry itself as a social institution participates in these loops through diagnostic labelling, discursive practices, and modes of social control that may aggravate or ameliorate suffering (200,201). The types of problems included within the purview of psychiatry, the kinds of explanation and interventions used, and the larger context of practice are all part of the dynamic system that shapes experience and behavior. --- 4E cognitive science as a path to multilevel integration Contemporary 4E cognitive science points to ways to conceptually integrate multiple dynamic levels of organizational complexity that involve neurobiological, social, cultural, and environmental contexts across spatio-temporal scales (187,(202)(203)(204)(205). The 4E cognitive science approach argues that cognitive processes are embodied, embedded in social contexts, and involve enactments that extend into the world. Embodiment refers to the ways in which the body provides a scaffolding for cognition and experience. 7 Enactment emphasizes that embodied experience emerges through ongoing cycles of action and perception that engage the environment. Cognition serves adaptation,

While contemporary psychiatry seeks the mechanisms of mental disorders in neurobiology, mental health problems clearly depend on developmental processes of learning and adaptation through ongoing interactions with the social environment. Symptoms or disorders emerge in specific social contexts and involve predicaments that cannot be fully characterized in terms of brain function but require a larger social-ecological view. Causal processes that result in mental health problems can begin anywhere within the extended system of body-person-environment. In particular, individuals' narrative self-construal, culturally mediated interpretations of symptoms and coping strategies as well as the responses of others in the social world contribute to the mechanisms of mental disorders, illness experience, and recovery. In this paper, we outline the conceptual basis and practical implications of a hierarchical ecosocial systems view for an integrative approach to psychiatric theory and practice. The culturalecosocial systems view we propose understands mind, brain and person as situated in the social world and as constituted by cultural and self-reflexive processes. This view can be incorporated into a pragmatic approach to clinical assessment and case formulation that characterizes mechanisms of pathology and identifies targets for intervention.

dversity determine the causes and course of symptoms. Psychiatry itself as a social institution participates in these loops through diagnostic labelling, discursive practices, and modes of social control that may aggravate or ameliorate suffering (200,201). The types of problems included within the purview of psychiatry, the kinds of explanation and interventions used, and the larger context of practice are all part of the dynamic system that shapes experience and behavior. --- 4E cognitive science as a path to multilevel integration Contemporary 4E cognitive science points to ways to conceptually integrate multiple dynamic levels of organizational complexity that involve neurobiological, social, cultural, and environmental contexts across spatio-temporal scales (187,(202)(203)(204)(205). The 4E cognitive science approach argues that cognitive processes are embodied, embedded in social contexts, and involve enactments that extend into the world. Embodiment refers to the ways in which the body provides a scaffolding for cognition and experience. 7 Enactment emphasizes that embodied experience emerges through ongoing cycles of action and perception that engage the environment. Cognition serves adaptation, and a changing environment requires action to maintain the body and the person in a healthy, functional state (208). Human adaptive niches are cooperatively constructed. Action and experience therefore are embedded in social-cultural contexts. The action-perception cycles of cognition extend beyond the body to engage with the material and cultural affordances of a local niche and larger social systems. From a 4E perspective, both the experience and the mechanisms of health and mental disorders can be approached in terms of individuals' dynamic engagement with the social world. Dynamic engagement with the social world requires constant adaptation and resource optimization. The concept of allostasis, which refers to the ways in which organisms anticipate and adapt to challenges, focuses on the function of physiological and biobehavioral systems of stress response and regulation (208,209). Allostasis involves the organism's capacity to allocate resources to maintain an adaptive balance between coping and recovery in response to adverse conditions and events. This involves both internal physiological processes and behavioral strategies based on appraisal of challenges and available resources for coping (210). When allostatic regulation is insufficient, various forms of stress-related dysfunction can result from has been described as 'allostatic overload' (211). The processes involved in allostatic regulation can be viewed from an enactive perspective as ongoing cycles of action-perception (212). They can also be modelled as Bayesian processes of active inference, in which the organism predicts and acts on the environment to ensure its own stability (108). These cycles occur internally through interoception and physiological regulation of the internal milieu and externally through behaviors that act on the body and the environment (213,214). Cycles of action-perception also underlie our sense of agency both in terms of the sense of volition and control (215), and the wider sense of being able to change our social circumstances (216)(217)(218). The action-perception cycles that are constitutive of agency and subjectivity emerge in and are maintained by social-cultural contexts that involve other people in dyads or couples, families, neighborhoods and communities, as well as larger social networks and institutions (219). These larger ecological domains contribute to higher-order goals and plans. Problems in self-regulation and adaptation can originate at any level in this system, with potential repercussions throughout. Hierarchical organization of goals is part of healthy functioning and certain forms of psychopathology may result when stress or allostatic overload disrupts this organization (220). Healing practices, therapies and treatment interventions can work to restore allostatic function where it has been disrupted. The overall aim of allostasis is to adjust regulatory systems to maintain the health, survival and reproductive fitness of the individual. More proximally, this includes responding to the challenges and demands of a social niche in ways that fit local cultural norms, roles and expectations. This may involve changing perceptions (learning new ways to attend to and interpret sensations from the body or the environment), taking new actions (enlarging the repertoire of behaviors and changing plans and priorities), or re-establishing links between action and perception that have been disconnected (providing feedback from outcomes that can guide recursive goal setting). Both internal changes and actions on the world can participate in the same adaptive cycles. The 4E approach can be readily extended to include the essential functions of language in human adaptation (221). Humans are language animals (113), inhabiting a world that is comprised not only of physical arrangements but saturated with linguistically mediated meanings, which provide the content of social norms and conventions as well as the scaffolding for the construction of a narrative self. The narrative practice hypothesis focuses on how this linguistic capacity emerges developmentally through culturally prescribed practices of self-narration, giving rise to folk psychology with its grammar of motives, plans and intentions that are employed to organize memory and action, articulate individual goals, and offered to others as reasons and explanations for one's behavior (222). Linguistic capacities allow regulation of systems that are organized in terms of physical dynamics because narrative construals of self and context organize, constrain and modify lower-level action plans both within individual cognition and in communicative interactions with others. Language is selfreferential and recursive and, through metaphor and narrative, is used by individuals and groups to construct novel multilevel hierarchies that regulate complex cognition and behavior. This is a key facet of the ways that culture permeates human cognition and functioning. Of course, language and culture reach deeper to reshape cognition, perception and action in ways that are nonconscious, implicit and automatic (167,207,223,224). Throughout the lifespan, culture shapes the human nervous system, allowing us to navigate socially constructed environments, engage in cooperative activities, and pursue our goals through embodied knowledge, skills, habits and dispositions (225). But much of culture remains outside the individual, distributed among others with specific expertise, residing in relationships, reproduced in institutions or practices, and present in social niches that provide cultural affordances for action and perception (226). These cultural affordances are part of the extended context on which human cognition and adaptation depend. Central to this context are interactions with other people, texts, and institutions. We rely on these interactions in local niches and relationships or larger networks to scaffold cognition, guide behavior and augment our capacities by "thinking through other minds"-whether in ongoing cooperative interactions with others or by consulting the vast archives of human knowledge and experience (44). In summary, current elaborations of 4E cognitive science offer an account of human function in dynamical systems terms as embodied (coupling bodily physiology and experience), enacted (involving sensorimotor loops that give rise to agency), embedded (context sensitive), and extended into the environment (dependent on cultural affordances). By tracking the ways that processes of organismic selfregulation and experiential learning emerge from ongoing cycles of interaction between the individual and the social-cultural environment, this framework can integrate physiology, cognitive processes, including individual agency and self-construal, and participation in cooperative meaning-making. This allows us to recast basic processes of symptom production, distress, coping and adaptation as well as the response to interventions in terms of multilevel dynamical systems. This systemic view opens the way toward a conceptual approach that considers how the co-constituted systems of body, mind and person are in transaction with larger interpersonal, social and cultural systems. --- Integrative case formulation Comprehensive diagnosis and treatment in psychiatry requires addressing pathology in all its dimensions: biological, psychological, social, cultural, and environmental. Integrating these into causal explanations of particular types of problems remains a challenge for psychiatric theory and practice (126,227). Approaching these multiple forms of explanations as independent or even incommensurable ignores the obvious ways in which processes at multiple levels not only affect but mediate each other. An ecosystemic approach to integration aims to identify multiple causal processes or mechanisms within and between levels of organization and articulate their connections in an overarching system. Advancing integrative case formulation requires approaching the patient as embodied and embedded in an ecosocial niche that presents an array of inter-related social determinants of health with differential constraining and enabling opportunities. The same niche also provides models for self-understanding, values, aspirations, and afflictions that shape experience, adaptation, coping, and help-seeking behavior, as well as access to services, educational and vocational opportunities, and other resources. Individuals' responses to adversity, symptoms or disorders, and modes of recovery will be influenced by the norms, expectations, and constraints of the sociocultural contexts they inhabit. To illustrate how this integrative perspective works in clinical practice, consider the following case vignette8: A 30-year-old woman presents to a mental health clinic with a self-diagnosis of depression. On inquiry, she reports feelings of emptiness, worthlessness, and guilt, as well as irritability, restlessness, rumination, difficulty concentrating, indecisiveness, early awakening, and fatigue over the past 6 months. Most recently, she has had increasing loss of interest and pleasure in ordinary activities and social isolation, as well as thoughts of death. She has done some online research and comes to the clinic asking for laboratory tests to confirm her diagnosis and determine the best treatment. She recently read a blog that mentioned novel research findings on the use of brain imaging and pharmacogenetics in personalized treatment for depression and presents the clinician with a list of private labs that offer this service. On further discussion, she reports that she lost her job three months ago and feels deep humiliation. She also mentions having difficulties in her relationship with her partner, saying that they are "going through a rough patch. " She explains that she feels anxious and out of control and at times fears that she is "losing my mind. " She is prescribed an SSRI antidepressant and experiences some lessening of her symptoms over the next few weeks, but does not feel any return of sexual interest, which adds to her worries about her relationship. As is increasingly common in mental health care, the person in the vignette presents clinically with a self-diagnosis of depression and, in this case, expects treatment with medication for what she views as a brainbased disorder. She also has ongoing social stressors that may be both causes and consequences of her mental state. How she interprets her symptoms and her feelings of anxiety, hopelessness, humiliation, guilt or shame will affect both her behavioral and neurophysiological response to the predicaments of job loss and relationship strain. In addition to temperamental traits or constitutional predispositions and the neurobiology of mood regulation (228), a complex interaction of embodied processes-shaped by previous illness experience, life events, and the response of others-add reinforcing or attenuating loops that further complicate the system dynamics that underlie symptoms and distress. A clinically effective approach to explain and treat distress therefore must go beyond neural correlates and biomarkers to consider individual variations in phenomenology and lived experience (229,230), developmental processes (231,232), symptom trajectories (233,234), and socio-cultural dynamics, which depend on social structure, institutions and practices, as well as cultural systems of meaning (218,235,236). In the case of the patient in the vignette, the causal mechanisms of anxiety, demoralization and depression can (and likely do) start at many different points in the network depicted in Figure 1. Additionally, each of these processes can interact with potentially reinforcing or compensatory feedback loops. These dynamics are important for adequately characterizing the nature of the problem, its Many of the links shown in Figure 1 are mediated by personal, social and cultural processes of meaning making. These involve bodily and discursive practices as depicted in Figure 2. While physical stressors may have direct effects on physiology and elicit responses, based on past experience, that occur outside of awareness, the impact of stressors also depends on individuals' perception and interpretation of the event. This involves embodied and enactive processes of meaning-making that build on developmental experiences and draw from cultural resources (204). The process of meaning-making includes the person's appraisal of the level of threat, their coping skills and resources, and the potential consequences-that is, "what's at stake" for the individual and others in their social world (238). For example, while job loss is likely to be a stressor for most people, the degree of perceived stress and ability to cope will depend on contextual factors including the personal and cultural meanings of one's occupation and of unemployment, current economic resources, social supports and mobility. Shame and humiliation follow from experiences of loss of social status and failure in performing according to social norms (239). The experience of humiliation in response to job loss depends on its timing (e.g., family just moved for the job or has had other resource depleting stressors), social position, roles, norms and expectations (e.g., father expects to be a breadwinner). Social validation of perceived stress can also contribute to self-regulation and reduction of perceived stress through process of feeling understood, supported and protected, as well as helping the individual to shift perspectives, mobilize problem solving strategies, and access stress-reducing resources. Perceived stress can prompt multiple maladaptive behaviors that feedback in loops that lead to resource depletion. For example, drug consumption for symptom control, relaxation and or escape can lead to emotional lability and irritability that challenge relationships. In favorable constellations however, response to perceived stress may lead one to develop new skills or positive schemas, overcome engrained biases, rescript self-understanding narratives, expand one's affordances, deepen social relationships and improve coping. In the ecosocial systems view, interpersonal dynamics, work stress, gender discrimination, and cultural knowledge and practices for dealing with distress-all of which depend on or reside primarily in social interactions-may contribute to the patient's distress, coping strategies and process of recovery. Applying an integrative perspective in case formulation requires considering how these processes unfold over time in the individual's life trajectory. Moreover, the processes related to each of these levels and dimensions interact in ways that can give rise to feedback loops that exacerbate symptoms and result in a Ecosystemic Embedding of Depressive Symptoms. The figure illustrates some of the many links between symptoms, processes and experience that constitute the ecosocial system of the patient described in the vignette. The arrows represent causal influences mediated by diverse mechanisms. Closed loops can give rise to feedback amplification, resulting in vicious cycles of symptom exacerbation or, when regulatory mechanisms are sufficient, can lead to allostatic changes that contribute to resilience and recovery. For the links marked with asterisks, the mechanisms of influence depend on nonverbal and linguistic communication through embodied and enactive loops that give rise to intersubjectivity, positionality and ongoing negotiations of meaning as depicted in Figure 2. Based in part on (237). Gómez-Carrillo and Kirmayer depressive disorder or other syndrome, which may then be maintained through similar looping mechanisms (237). These loops are not only internal to the brain and its circuits but extend beyond the body to social interactions with other people and social institutions-all of which affect the development and course of psychiatric disorders. The cultural-ecosocial approach is fundamentally relational. The relationships it considers involve material, informational and symbolic-communicational interactions between the individual and the environment. These relationships can be mapped by causal loop diagrams (CLD) that aim to capture the links between observable processes (240). These maps can be used to develop formal quantitative models to reveal dynamics and test the potential impact of interventions, including changes in the configuration of systems-e.g., by altering individual biology or cognition, family interactions, health care systems or other social contingencies (241,242). In the ecosocial view, humans are embedded in and dependent on culturally constructed environments that include physical arrangements as well as a web of relationships with other people and social institutions. The 4E perspective insists that interactions with the environment are part of the dynamics that constitute the individual. In human ecology, however, the distinctions between individual and environment are phenomenologically, psychologically, morally and politically important. Hence, drawing the boundary between 'inside' and 'outside' (organism and environment or system and subsystem) varies with the clinical question and the way we locate the relevant dynamics (243). There can be principled and practical reasons for drawing a boundary in a particular way both because it highlights crucial dynamics and constitutes a useful way to organize case formulation and guide intervention. These reasons may include the system's topology, the feasibility of specific interventions, and the ethical imperative to privilege the patient's perspective (244). --- Integrating the patient's self-understanding A key element in an ecosocial systemic approach is recognizing the role that the person's own understanding of and response to symptoms and suffering play in the dynamics of mental disorders, coping, help-seeking, treatment response and recovery. In the case vignette presented in the previous section,the patient's self-diagnosis and explanatory model of her symptoms follow closely from the prevailing brain-centric model of depression widely disseminated in popular culture. This model portrays depression as a condition related to specific neurotransmitters and explains the efficacy of medications by their effects on corresponding receptor sites. More recent versions of this explanatory model go beyond synaptic mechanisms to consider brain circuitry (245)(246)(247). Other patients may present explanations that draw from sociomoral or religious understandings of suffering and view illness as a consequence of moral transgression or failing (248). These modes of explanation and attributions influence ways of coping and help-seeking but they may also participate in the vicious circles that aggravate dysphoria, self-deprecation, social withdrawal, and other symptoms of depression (249). Embodied and Enactive Processes of Meaning Making. The figure outlines the cyclical processes of embodiment and enactment that give rise to meaning and experience. Experience emerges through developmental processes and engagement with others in particular social-cultural contexts. There is two-way traffic between bodily processes and individual experience mediated by cognitive processes of metaphoric thinking, imagery and imagination. Similarly, there is traffic between experience and social discourse mediated by interpersonal communication and narrative practices. All of this occurs in a field of cultural affordances provided by local niches and larger social contexts (Adapted from The patient's illness narrative, which emerges in dialogue with available cultural models and in clinical encounters, also shapes the process of meaning-making and illness experience (238). The models used by clinicians -which borrow from both technical literature and dominant cultural narratives -also shape patients' experience and expectations (173). In this case vignette, the patient adopted a simple biological model of depression even before speaking to the doctor, setting aside her challenges of job loss and relationship problems as secondary issues. In so doing, she focused her expectations in consulting the clinician on receiving a specific medication. While this fits squarely with psychiatrists' competence, it may require negotiation, because her self-diagnosis may not be accurate and her requested treatment may not be appropriate, and, even if it does address an important facet of her current problem, medication may not be sufficient to resolve other aspects of her predicament (250,251). The effects of adopting a neurobiological explanation go beyond a narrow focus for clinical assessment and treatment to also influence the patient's sense of self-efficacy and participation in the process of recovery as well as broader features of her identity. A simplified, brain-centric model of depression makes antidepressant prescription seem a straightforward, necessary, and sufficient clinical response. Of course, beyond pharmacogenetics, kinetics, and dynamics, our mechanistic knowledge of drug action remains limited (252). Antidepressant treatment may have different effectiveness based on the individuals' expectation of efficacy (253) or their socioeconomic status (254), requiring the clinician to consider the interaction of the type of treatment and the patient's context when collaboratively designing a care plan (255). Moreover, prescription is inevitably a social and symbolic act, and taking medication has meaning and consequences for psychological selfregulation and social identity (256,257). Rose (258) has drawn attention to the ways that biomedical diagnosis and treatment of mental disorders lead to narratives of "neurochemical selves" with consequences for individual coping as well as for mental health policy and practice. There is increasing recognition that good practice in psychopharmacology requires paying attention to the personal and cultural meanings of medication and patients' own values and priorities (259). A cultural-ecosocial view can inform existing approaches to shared decision making and collaborative prescribing or deprescribing of medication (255). --- An ecosocial systems approach to person-centered clinical practice Psychiatric practice employs multiple ways of knowing that have been characterized as verstehen (understanding), erklären (explaining) and einfühlung (empathic, embodied co-presence/being/knowing) (260). These ways of knowing have different epistemic bases and constraints and are sometimes in tension, conflict or competition. In contemporary psychiatry, this tension is seen between the divergent approaches of precision psychiatry (which characterizes the person in terms of biological parameters) and person-centered psychiatry (which emphasizes experience, values and context) (261,262). Although advocates of each approach superficially acknowledge the other, in practice their respective research programs and modes of implementation reflect the persistence of an underlying dualistic ontology (129,218). Bringing erklären, verstehen, and einfühlung together in clinical formulation means integrating explanatory models and mechanisms across levels, including molecular, physiological, neural circuitry, cognitive, and social. Including the social level requires knowledge of social and cultural history and current context as well as biographical trajectories. Because our institutions and practices are embedded in these same contexts, a social-cultural perspective requires self-reflective consideration of the clinician's positionality and interaction with the patient and others in the co-construction of clinical narratives (260). The cultural-ecosocial systems approach offers a frame that can encompass these dimensions of psychiatric practice through a dialogical process of meaning-making that recognizes culture and context. Human ecological niches are fundamentally social-with socially constructed contexts and relationship providing the essential matrix of development from inception-and cultural, with shared meanings, values and practices shaping cognition and experience across the lifespan. The notion of ecosystem builds on work in ecological systems theory in developmental psychology (68), which emphasizes the embedding of the individual in multiple, nested environmental contexts, defined by socio-relational and spatio-temporal scale and composition to include: micro (immediate family and friends, community and work-school setting); meso or exo (neighborhoods, wider networks, and larger community); and macro (society, nation, transnational) contexts. (See: Table 1), The idea of a niche highlights the interactive and dynamic nature of such sociocultural embedding. Social context, structural, economic and political forces affect individuals and groups differentially as a result of individual and collective past histories, biology, and current positionality (263). To unpack the notion of niche in a way that can serve a personcentered clinical approach, the ecosocial systems view needs to consider the intersections and interactions across at least four overlapping domains: (1) lifespan developmental history; (2) social structure and positioning; (3) cultural meaning, norms, values and affordances; and (4) individual biography and self-understanding (which draws selectively from each of the other domains). These domains can provide a temporal dimension to clinical formulation that points both to adaptive challenges and resources for helping, healing and recovery. Efforts to develop models that incorporate social context and lived experience are underway, but they face multiple obstacles, including lack of collection of data representative of population variability and high levels of context dependence as well as ethical and pragmatic issues related to the use of such data (264). We need better conceptual, research and clinical tools to characterize niches-their demands, affordances, and constraints as well as their embedding in larger ecosystems (6). The theory of syndemics provides one approach to exploring the multilevel interactions that give rise to mental health problems (265,266). While the notion of niche points to the immediate environment that an individual inhabits, in reality, human niches are subsystems of larger social systems. An ecological view encourages us to examine this larger network of relationships and how they interface with local niches. It is a virtue of the ecological perspective that it allows us to think systematically about the relationships between our most proximal and intimate relational networks and the larger networks with which we are coupled. The nature of this coupling depends on local arrangements and interpersonal interactions, which are extended by population migration as well as information and communication technologies that allow connections with distant others but that also create virtual environments that we increasingly inhabit (267,268). In the current moment, relationships on the planetary scale are increasingly present and consequential in the lives of individuals through the impacts of climate change (269,270). These interactions occur in material ways, but they are also present in self concepts, imagination and orientation toward the future with significant mental health impacts. True to its name, an ecosocial view, encourages us to think about mental health as dependent on these wider networks and modes of interdependence. Coming to terms with the impact of our changing environments requires considering not only strategies for individual adaptation, but the larger, social structural arrangements that account for global disparities and that constrain the options of individuals and groups across the globe (271,272). Ultimately, mental health theory and practice must consider not only the private challenges of individuals, but the larger dilemmas faced by our species and the planet we share with others (273). --- Conclusion Although psychiatry conventionally locates mental health problems in the individual, systems thinking encourages to see the ways in which health and the wide range of problems seen in clinical settings arise from interactions at multiple levels from the biological to the cognitive and social. Recognizing patients' agency and restoring their health requires that clinical care consider the range of systemic processes that contribute to suffering and impairment (274). Addressing problems that derive from social structure may require interventions that go beyond individual clinical care to include advocacy and social-network interventions. Advocacy is not limited to efforts to change policy and institutional practices but includes actions that aim to counter oppressive circumstances and create habitable environments and niches for individuals (275). Efforts to provide multilevel systems explanations of health problems are often challenged as "too complex" for practical application. Systems dynamics may be difficult to think through and require specific training to apply. Complex systems can exhibit counterintuitive properties, but qualitative understanding is often sufficient to guide practice (276)(277)(278)(279). Quantitative models of specific problems could allow clinicians to examine the effects of potential interventions on system dynamics to guide treatment and predict outcomes. Crucially, these models can include clinician-patient interaction and other social processes as part of the symptom network. Innovative computational methods can capture multilevel system dynamics if the relevant data are collected (264). The resultant models could be used as decision tools or used by clinicians and patients to foster mutual understanding and motivate interventions. The models we offer to patients are themselves interventions that may guide selfreflection and elicit new behaviors. They may also function as selffulfilling explanations that foreclose the search for better answers. How this plays out depends on the ability of the clinician to apply dynamical systems models while closely attending to the patient's experience so that the model can be refined and care remains patient-centered. The application of dynamical systems models in psychiatry, though actively pursued for decades, has been slow to advance and has had limited uptake. There are several likely reasons for this, including that the adoption of systems thinking has been hampered by (i) continued investment in reductionist models because they are amenable to study by common scientific methodologies; (ii) the limitations of clinical decision making, which make it hard to incorporate complexity and interaction effects; and (iii) economic and political interests that favor short-term treatment and pharmacological interventions rather than approaches that challenge entrenched systems. However, new computational modelling methods that can be implemented in clinical settings to support patient education and real-time decision making offer the hope of significant progress. The challenges associated with complexity reflect the real-world dynamics of human problems (280,281). Recognizing this complexity should urge on us humility and the need to frequently recalibrate our clinical response to respond to patients' experience. It underscores the need for idiographic methods of case formulation, which may include characterizing networks of relationships among symptoms and related biological, cognitive, and social processes (194,282). Finally, it points to the importance of self-reflexivity, in which clinicians interrogate their own assumptions and practices to rethink case formulations and potential interventions. The cultural-ecosocial view includes practitioners, clinical settings, health care systems and the local and international institutions of psychiatry itself -both as material and discursive practices-as part of the systems in which patients and practitioners are embedded and which offer them affordances, norms and constraints. These need to be factored into practice in general and into the formulation of specific cases. A literature in critical psychiatry has considered some of the ways in which psychiatry colludes with larger structures of oppression (201,283). This is more likely to occur when psychiatric practice is narrowly conceived as the identification and treatment of discrete disorders without attention to patients' lived experience, values, and lifeworlds as well as to practitioners' tacit assumptions. By giving an explicit place to the meaning-making process in clinical encounters as well as in institutional and wider social contexts, a cultural-ecological systems view opens the door to more self-reflective and critical thinking that can uncover power dynamics and counter potentially oppressive practices. An ecosocial systems view offers a way for clinicians to organize the multiple explanatory models needed to capture the complexity and heterogeneity of psychiatric disorders and illness experience. Based on a view of psychiatric disorders as involving complex system dynamics, an ecosocial systems approach allows clinicians to use multiple languages of description to assess processes within and across levels of organization of an overarching ecology of mind and to prioritize those that offer the greatest therapeutic leverage and optimal use of resources for person-centered practice. --- Data availability statement The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author. --- Author contributions AG-C and LK contributed equally to conceptualizing the manuscript. AG-C wrote the first draft. LK wrote sections of the manuscript. All authors contributed to the article and approved the submitted version. --- Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. --- Publisher's note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

While contemporary psychiatry seeks the mechanisms of mental disorders in neurobiology, mental health problems clearly depend on developmental processes of learning and adaptation through ongoing interactions with the social environment. Symptoms or disorders emerge in specific social contexts and involve predicaments that cannot be fully characterized in terms of brain function but require a larger social-ecological view. Causal processes that result in mental health problems can begin anywhere within the extended system of body-person-environment. In particular, individuals' narrative self-construal, culturally mediated interpretations of symptoms and coping strategies as well as the responses of others in the social world contribute to the mechanisms of mental disorders, illness experience, and recovery. In this paper, we outline the conceptual basis and practical implications of a hierarchical ecosocial systems view for an integrative approach to psychiatric theory and practice. The culturalecosocial systems view we propose understands mind, brain and person as situated in the social world and as constituted by cultural and self-reflexive processes. This view can be incorporated into a pragmatic approach to clinical assessment and case formulation that characterizes mechanisms of pathology and identifies targets for intervention.

norms, and environmental knowledge 10,11, and these goals may not automatically align with plans developed from an ecological perspective. In order to respect this diversity, ecologists must proactively develop transdisciplinary partnerships 7,12,13 with sociologists, landscape architects, urban planners, and economists. With these partners, projects can connect with community members through iterative listening sessions, neighborhood surveys, focus groups, and planning meetings that focus on communitybuilding and long-term conservation gains 1,7,14. While we sought partnerships during project development, we found that the scale of our study was too large for our team to effectively engage at the community level. If we had worked in fewer neighborhoods and prioritized connecting with neighborhood-scale organizations instead of city-wide governance, we would have better understood the concerns of neighbors. Likewise, we would have benefited from sociological expertise when assessing and responding to community feedback, instead of only when identifying best practices for our proposal. Furthermore, our focus on economic feasibility and large-scale implementation necessitated a simplistic habitat design with minimal management costs, but this proved unrealistic. Working closer with landscape architects would have helped us better anticipate and resolve neighbors' aesthetic and safety concerns in our initial habitat design rather than through iterative modifications. Urban conservation projects that effectively use neighborhoodscale organizational partnerships and invite community members to participate in goal setting and design are better primed for a b c d e f Fig. 1 Urban greening projects should anticipate challenges and plan to work collaboratively with stakeholders to address them. Post-industrial cities such as Cleveland, Ohio, USA face the significant challenge of managing thousands of vacant lots (i.e. parcels where pre-existing structures have been torn down and replaced by minimally-managed vegetation) (a). Although typically viewed as blight, these sites do offer opportunities to conserve urban biodiversity. In 2014, the Gardiner Laboratory established 64 conservation habits on vacant land across the city of Cleveland, including 32 sites seeded with native perennial wildflowers (b). Our cost to establish the 32 pocket prairies was approximately $2500 per site and included soil preparation, seeding, invasive plant management, and installing 'cues to care' 11 such as fencing and mulching. Maintaining cues to care required substantial time and financial investment. Trash, furniture, appliances, and other refuse are frequently dumped into urban lots and must be removed (c). Vandalism to signs, fencing, and vegetation is also common, especially in the springtime when native plantings have not bloomed yet and sites can appear weedy and unkempt (d). Similar aesthetic concerns can happen during the winter. Thus, it is imperative to work with local stakeholders to identify what design modifications can indicate year-round investment into a conservation site. For instance, greenspace managers at the Sunflower+Project: STL, in St. Louis, Missouri, partnered with local elementary schools to paint sustainability flags which hung above their over-wintering sunflower fields (e). However, even with substantial investment, vacant lot ecosystem management can remain a controversial endeavor. Although many residents view rain gardens established though a multi-million USD investment by the Northeast Ohio Regional Sewer District as valuable storm water management, others see a poor use of funds that raises health and safety concerns 4 (f). Thus, we must work with urban residents and municipal governance as community developers to meet the needs of the diverse human ecosystem. Photograph E courtesy Richard Reilly. success. For example, the Burnham Wildlife Corridor in Chicago, Illinois partnered with the Field Museum and the Chicago Park District and synthesized 20 years of participatory action research on community perspectives to create their "Roots and Routes" initiative 1. This program successfully aligned their conservation goals with residents' perspectives on greening to simultaneously develop migratory bird habitat, gathering spaces for communities, and opportunities for youth engagement and employment 1. In part, we attribute "Roots and Routes" success to their design-competition approach that allowed community groups to determine their own project goals and habitat plans; this should inspire future projects to creatively incorporate community perspectives during grant development. --- Project sustainability relies on community relationships Co-creating a greenspace is only the beginning; conservation practitioners must also communicate progress and maintain community member's trust throughout a project's duration 15. Despite residents' involvement in the co-design process, feelings of "bait and switch" can arise if a developing habitat aesthetically diverges from their expectations. For instance, non-native weedy vegetation may become more abundant within a planting than anticipated. Thus, frequent discussions of all possible or transitory site outcomes, including visual representations of a habitat's vegetation 9,14, can help avoid feelings of contention or disinvestment. Likewise, research tools can be misunderstood and cause concern if not effectively described. For example, neighbors have expressed apprehensions that our native bee traps were releasing stinging insects when in fact they removed the insects for further study. This confusion could have been avoided by better communication at the project's onset and throughout continued interactions with residents. To effectively engage a large and diverse urban community, researchers must evaluate multiple options to share their activities and findings 14,16. We created a project website, educational video, and social media presence; these have been successful in communicating with other researchers and the media but have largely failed to reach residents. We found that one-on-one discussions of project aims, progress, and outcomes through daily interactions on-site or at community events were far more effective, but still did not reach all stakeholders. For example, our research activity occurred during the day, limiting interactions --- BOUNDARIES What transdisciplinary partners are critical throughout the project's life cycle? 7,12,13 What neighborhood norms for landscape design typify local communities? How do we reconcile these with novel designs developed through the co-creation process? 7,11 What is our plan for identifying, implementing, and iteratively adapting cues to care after receiving feedback? How do we ensure our adaptive responses occur in a timely manner? Is participation in data collection and analysis desired by community members? If so, how do we facilitate this? 20 To what extent do community members wish to participate in site preparation, installation, or maintenance? Should we consider hiring a community organization or businesses to coordinate project implementation or management? How do we promote community leadership and work with diverse stakeholders to co-create our project goals and objectives? 14,18 How can our research team build and maintain trust with neighbors? 15 How will we ensure that all project communications to the public and media outlets are consistent, accurate, and reach their targeted audience? 19 What methods are most effective to gather community knowledge and opinions throughout the project? How frequently should we solicit this feedback? 16 Who will serve as the liaison for community concerns? What governmental ordinances, regulations, permits, and safety considerations constrain our decision making, site selection, and timeline? 17 How will the greenspace be funded in the short (e.g. grant cycle) and long-term and are there restrictions on how funds can be used? What is our plan for addressing community concerns that exceed our grant budget? How can we achieve our project goals with cost-effective and scalable methods that enable future site development in other communities? 15 As individuals with varying degrees of environmental knowledge encounter the project, how can we effectively communicate its goals and merits? Are there on-site learning opportunities that could facilitate community understanding? What existing neighborhood organizations, faith communities, or governmental programs represent important social networks? Are there keystone individuals who foster these? How do we make connections? 1 a b --- Implementation Assessment Fig. 2 Guiding questions for ecologists planning future urban greening projects. A "for the city" paradigm for ecological research approaches urban conservation as an iterative community development process (a) for the benefit of urban residents and taxa of interest. Applying such frameworks can be difficult for scientists with disciplinary knowledge but little expertise working with a diverse set of stakeholders. We recognize that praxis often falls short of aspirational theory and provide the following set of questions and references (b) for consideration by practitioners embarking on new projects 1,7,[11][12][13][14][15][16][17][18][19][20]. As urban conservation is complicated and context dependent, this is not an exhaustive list and references often apply to multiple questions. Moreover, we emphasize that both community development and urban conservation are long-term endeavors, not activities bounded by a research grant, and that iterative adaptions are critical to achieving positive outcomes. with those who were away at work, and a high turnover rate in housing occupancy created an influx of new residents unfamiliar with the project. Also, residents may be more comfortable interacting with neighbors or local organizations rather than visiting researchers. To address these issues, scientists should consider pre-existing channels of communication (e.g. neighborhood watch groups, religious organizations, community centers) that are self-identified by community leaders. Building community relationships requires more than disseminating information; researchers need to actively solicit community opinions in order to gauge and address needs. Regular polling of community opinions through focus groups or surveys can inform habitat management to help resolve community concerns such as aesthetics or perceived safety. While an ecologist may see diverse native plants flourishing, dense and tall plantings can inspire fear of criminal activity 17 and community members may consider such habitats as eye-sores indistinguishable from abandoned properties 11. Such issues can be mitigated through "cues to care"-the physical signs of intention and upkeep 11 advocated by design professionals. These cues can indicate a greenspace has purpose, help combat negative perceptions, and illustrate community consideration 11. Common practices such as signage, neatly-mown borders, fences, and/or mulched flower beds around a conservation site can convey a site's purpose and contribute to community approval 9,11. For example, abundant signage and neatly-mown edges were noted as significant factors promoting public support for urban meadows in Bedford and Luton, UK 9. We employed similar cues to care and framed our pocket prairie habitats with a mown border, fence, and mulched roadside edge (Fig. 1b). Yet, we received complaints from residents who did not perceive our mown borders as intentional and assumed we had abandoned our mowing efforts. Concern was also expressed that the mulch was a health hazard as stray cats might use it as litter. This illustrates how widely recommended cues to care are not effective in all settings and failure to engage residents in management planning may result in confusion or elicit unanticipated, negative feedback. Conversely, if residents are involved in determining cues to care, creative solutions generating greater satisfaction can be found. For example, hand-painted flags designed by elementary school students were an effective cue to care for the off-season within a sunflower planting in St. Louis, Missouri 5 (Fig. 1f). It is important for community leaders, scientists, and neighbors to recognize the difficulty in reconciling a community's diverging opinions of greenspace goals. Even with open communication, projects will face challenges in meeting community expectations. For instance, some Cleveland, OH residents prefer the tidy appearance of fabric flowers over the living vegetation of a habitat planting. After 4 years, we are still trying to develop a strategy to meet this concern. Meanwhile, we have also received many positive comments, with residents enjoying the color of our plantings, asking to pick flowers for bouquets, or declaring their general support for helping declining bees. We highlight these variable responses as both precaution and encouragement. It is unlikely that urban conservation sites will garner universal public support 2, but iterative assessments and modifications of a site's management or design can ameliorate some community concerns and shift how greenspace is viewed and valued long-term. --- Final remarks Balancing the diverse needs of human and ecological systems is complicated, context-dependent, and relies upon partnership, detailed planning, and continued community engagement. Such an approach is critical to avoid driving a wedge between academics and urban residents, who may perceive researchers as outsiders charging in to "save" or "experiment on" inner-city neighborhoods. Although we are still learning how best to approach urban conservation, we recognize that connecting and co-creating with a broad group of stakeholders is a critical first step. Likewise, open lines of communication and timely management adjustments can help an urban greening project succeed. It is our hope that by approaching urban conservation as a community development process, we can collectively be better equipped to serve people and our taxa of interest. --- Author contributions K.J.T. and M.M.G. conceptualized, co-wrote, and edited the article and figures. --- Competing interests The authors declare no competing interests.

As investment in urban conservation grows, researchers must balance the needs of residents and conservation targets. We discuss some of the challenges we have encountered and the importance of taking a transdisciplinary approach informed by design and social knowledge. Urban greenspaces are increasingly considered as conservation habitats. In particular, vacant lots offer valuable opportunities (Fig. 1a). Vacant lots can be transformed into urban farms, rain gardens, and "pocket prairies" (Fig. 1b) in order to conserve biodiversity, deliver ecosystem services, and improve the equitable distribution of high-quality living conditions amongst city residents [1][2][3][4][5] . Urban conservation models champion an integrated social and ecological approach in creating greenspaces to account for the combination of biophysical, socioeconomic, and cultural factors 6-8 that shape city ecosystems. However, balancing socioecological theory and praxis is difficult 4,9 , and strategies for implementation are not universally applicable. Our insights are informed by a decade studying the ecology of vacant land within Cleveland, Ohio, USA, a post-industrial city that currently encompasses >27,000 vacant lots. In 2012, we received funding to evaluate eight economically-feasible strategies to manage vacant land, with the broad goals of improving habitat quality for arthropods, supporting ecosystem services 3 , and beautifying the city (Fig. 1). Despite familiarity with recognized socioecological frameworks, we struggled to apply recommendations to our project, particularly when establishing and maintaining native plants. We share our experiences of practical realities ecologists can face when attempting to follow best practices, and note the strategies employed by our team and others when implementing community-driven conservation (Fig. 2). Co-creating habitat goals sets a project up for success Paradigm shifts in urban ecology have emphasized the ethical responsibility of scientists to prioritize the city and its residents 8 by investing in community development and investigating questions relevant to human interests. Therefore, we collaborated with government officials at the Cleveland Land Bank, city council members, non-profit organizations, city planners, and other community leaders over a 12-month period to identify potential sites for our urban conservation project and finalize habitat plans. Following this, we canvased one city block surrounding each of the 64 vacant lots included in our study to discuss our research plans with residents and replace sites that drew irresolvable concerns. Despite endorsing a "for the city" paradigm 16 (Fig. 2), our project was not universally well received by residents and our habitat plantings experienced vandalism, dumping, and public criticism (Fig. 1c,d). In hindsight, we realize that we expected residents to tolerate our newly grant-funded project when we should have co-created project objectives with residents. Residents' goals for urban greenspaces may vary with their demographics, neighborhood cultural

Introduction The world's population is rapidly aging. Meanwhile, the number of senior residents living in nursing homes is increasing. Though nursing home residents often experience losses in physical health and functional independence, their psychosocial needs remain essential [1]. Social interaction is a critical factor that affects the quality of life of older adults. The quality and quantity of personal social relationships are linked to mental health and have implications for both morbidity and mortality [2]. Lack of social contact can leave individuals more vulnerable to disease, infarction, stroke, and the onset of Alzheimer's disease [3]. Nevertheless, social isolation is widespread among older adults in nursing homes, with senior residents' social connections being limited [4]. Even in nursing homes with high-quality care, residents often experience inactivity, motionlessness, and feelings of loneliness [5]. Entering a nursing home disrupts an individual's social world, significantly altering their relationships with family and friends [6]. According to socioemotional selectivity theory, a lifespan developmental perspective suggests that, as healthy individuals age, their social networks' breadth will be limited by the loss of friends [7]. Additionally, older adults are often disconnected from the mainstream social circle due to a lack of technology that resonates with them [8], and they are generally less likely to use technology than younger adults [9]. In response to the increasing need for social connection among older adults in nursing homes, we propose an Interactive Gallery, which has two primary objectives: to establish social connections between nursing home residents and citizens from the local community, and to facilitate social communication among the residents within the nursing home. To evaluate the effectiveness of the Interactive Gallery, we conducted direct behavior observations, including both baseline and intervention observations, as well as semistructured interviews in the field study. The results indicate that the Interactive Gallery had a positive impact on social interaction among the residents within the nursing home and stimulated their desire to share their lives with individuals outside of the facility. This paper's contributions and features are as follows. The social interaction of older adults constitutes a complex system that involves multiple stakeholders, including fellow residents, caregivers, members of the local community, etc. Previous research has primarily concentrated on strengthening the existing social relationships among nursing home residents. The Interactive Gallery aims to promote social interaction among nursing home residents and members of the local community, as well as between senior residents within the nursing home. We identify the social interaction system and behavior characteristics of nursing home residents. Based on a long-term deployment of the Interactive Gallery, we also outline strategies for promoting social interaction both among residents within the nursing home and between residents and local communities. --- Related Work 2.1. Theoretical Framing and Anthropological Studies of Social Interaction among Older Adults The concept of social isolation, particularly in the context of aging, refers to the inevitable loss of social and community ties [10]. Nicholson Jr has identified five broad causes of social isolation, including the absence of quality relationships, physical and psychological barriers, and socio-economic and environmental factors [11]. The literature has consistently shown that the quality and quantity of an individual's social relationships are related to their mental health and mortality [2,12]. Loneliness, which is associated with social isolation, has been defined in various ways, including the lack or perceived absence of satisfactory social relationships, a discrepancy between the relationships people have and the relationships they desire [13], and the amount of time spent alone due to a lack of social networks [14]. Directions of related anthropological studies are extensive. Gillian Harper Ice's study investigated how nursing home residents spent their day, and the conclusion was that older residents spent most of their time in their own rooms, sitting and alone [5]. Other researchers have further studied specific areas: Carstensen et al. have pointed out that older adults' social and emotional goals were directed towards strengthening existing emotionally fulfilling relationships rather than pursuing novel social partners [15]. Hilary Davis et al.'s study focused on the grandparent-grandchild relationship and e on intergenerational play in particular, identifying its pertinent features [16]. In recent years, more researchers have explored the question of how to apply social technologies to the easing of the social isolation of older adults. In this context, Alexis Hope et al. conducted interviews with older adults to understand their communication preferences for social media. They found that seniors articulate many concerns with social media, including the time required for participation, expectations of reciprocity, content irrelevance, and privacy [17]. In addition, some researchers use persuasion techniques to promote the social interaction of elderly people: John Paul et al.'s research focuses on developing a persuasive model employed by virtual agents for encouraging social interaction and lowering the risk of social isolation among older adults [18]. --- Applications of Older Adults' Social Interaction and Interactive Installations The applications for alleviating social isolation among older adults can be categorized along two main directions: (1) strengthening the connections between older adults and their existing social circles, primarily composed of family members; and (2) expanding older adults' social networks within the nursing home setting. For strengthening the connections, some designs attempt to combine an older adult's everyday objects with a display to enhance communication with family members. For example, Tsujita and Abowd designed SocialMedicineBox, a communication system for older adults using a medicine box [19]. Computer-mediated games, such as the adoption of a game console, such as a Wii, could play a role in supporting inter-generational interaction between older adults and youths [20]. Photographs are also seen as compelling media that facilitate communication: a paper-digital photo album could provide a focal point for communication between older adults and family members [21]. To expand the social circles, a system called StoryCube was developed to help residents of nursing homes to make connections through the sharing of stories and the expression of their identity [22]. Waycott, J. et al. investigated the role of digital content created by older adults in forging new relationships among strangers [8]. Their findings demonstrate that creating and sharing content provides opportunities for older adults to build new social connections within a small peer community. Interactive installations designed for nursing homes have received attention in the literature, with a particular focus on dementia residents. For instance, the authors of one study developed an interactive wall to assess the responses of nursing home residents with dementia to an interactive art installation, aiming to determine its effectiveness in reducing wandering behaviors [23]. In another study, a robot cat was designed as an artificial companion to provide interventions to older adults with dementia [24]. Moreover, the use of interactive robots has been explored to facilitate social interaction among older adults, such as the sociable robot "Paro," which was placed in a nursing home to encourage social interaction among residents [25]. Additionally, an interactive wall was developed to decrease the number of wandering behaviors of people with dementia [26]. --- Designing Technology for Older Adults to Enhance Health In recent years, there has been increasing interest in the design and development of technology for older adults. Ioana Iancu has provided a comprehensive literature review on the design principles and device features necessary to meet the needs of older people, serving as a valuable starting point for researchers and practitioners [27]. Giuliano Grossi's review focuses on frameworks involving computer vision and machine learning to promote elderly well-being, including cognitive, physical, emotional, and social aspects [28]. Achraf Ammar highlights the importance of innovative ICT-based solutions, such as the ICT-COVID-Companion, to improve physical and mental health among older adults, especially during pandemics. The study emphasizes the use of smart digital solutions, including emotional/social computing, open social platforms, interactive coaching, gamification, fitness trackers, and the internet of things, to provide safe personalized health surveillance [29]. Hacer Guner's study provides data from elderly citizens in Turkey, comparing their use and acceptance of ICT to that of younger adults. Their findings suggest that older adults have a positive attitude towards the use of technology, but still face some barriers, such as limited digital literacy and lack of awareness [30]. Keya Sen's study focuses on identifying how older adults can benefit from affordable and accessible technology use and how to tailor interventions to maximize their beneficial effect. Their study suggests that mobile technology-based applications not only help families stay connected but also link older adults to healthcare resources and encourage physical and mental well-being [31]. These studies highlight the importance of designing technology that meets the unique needs and preferences of older adults. By incorporating innovative approaches and smart digital solutions, technology can enhance physical, mental, and psychosocial health among older adults. Furthermore, these studies emphasize the need for tailored interventions that promote digital literacy and awareness among older adults to maximize the potential benefits of technology. --- Summary Existing research and applications have primarily focused on strengthening the preexisting social relationships of older adults, while giving less consideration to the involvement of individuals outside of the nursing home. However, both intimate and peripheral relationships in the social networks of older adults contribute to their social well-being, making a comprehensive study of their social interaction crucial. To obtain both subjective and objective information, our field study included inquiries from both elderly residents and caregivers. Furthermore, we aimed to develop a technology probe-an Interactive Gallery-to explore two dimensions of older adults' social interactions: how to establish connections between older adults and citizens from local communities, and how to promote social interaction among senior residents within the nursing home. --- Design Intervention --- The Current Social Interaction of Senior Residents In order to gain a comprehensive understanding of the social behavior patterns of older adults, we conducted interviews with both senior residents and their caregivers. A detailed description of the research process can be found in [32]. The following summarizes the main findings related to the current social interaction of senior residents. First, our findings indicate that senior residents tend to have alienated relationships with their fellow residents. Despite the fact that most older adults receive regular visits from their family members at least once a week, they still experience feelings of loneliness when their family members leave. The only chance for interaction among senior residents is during collective activities, which they engage in together. However, these activities are not sufficient to establish and maintain meaningful social connections. Moreover, most senior residents experience decreased mobility, which limits their ability to leave the nursing home without the company of others. As a result, the nursing home can be considered a relatively closed living environment for them. Overall, our study highlights the need for innovative interventions that can facilitate and enhance social interaction among senior residents. --- Interactive Gallery System Our design concept was developed based on the observation that senior residents frequently engage in sitting and watching outside of the nursing home through the window (Figure 1). Due to their limited opportunities for external interaction, we aimed to simulate the experience of being "outside" within the nursing home. We also drew inspiration from literature on the use of metaphor to reduce the barriers to technology adoption. A product designed with familiar metaphors has the potential to facilitate ease of use and reduce the learning curve associated with new technologies [33]. As a result, we adopted the metaphors of a gallery and postcard-sending, allowing for simple operations that the senior residents could utilize with minimal or no learning required. The Interactive Gallery system comprises two main components: scenery-collectors and a gallery-like interactive installation. Scenery-collectors are distributed to volunteers from local communities to share real-time scenery photos with senior residents. The gallerylike interactive installation is situated in the public space of the nursing home and allows senior residents to view and print scenery photos into postcards by simply pressing a button. These postcards can be used to communicate with the sharers or be shared with others. Additionally, the Interactive Gallery can also provide topics for the residents to communicate with each other. A use scenario of the Interactive Gallery system is presented in Figure 2: (1) The proposed system involves the distribution of scenery-collectors to citizens in local communities, which they can place wherever they choose to share scenic views with nursing home residents. The scenery-collectors are equipped with automatic cameras that capture the views at pre-determined intervals and transmit them to a gallerylike installation in the nursing home. (2) This installation displays the scenery photos and the addresses transferred from the scenery-collectors. (3) Senior residents can print photos as postcards by pressing a button on the installation. If the scenery photo reminds them of memories and experiences related to the photos, they could write them on the back of the postcards. (4) The postcards can be either exchanged with the senior residents or sent to the citizens who shared the view. To be specific, the Interactive Gallery detects the location and behavior of senior residents as they approach, triggering a dynamic 30-s slideshow of still pictures in a timelapse animation format. This allows elderly residents to observe delicate changes in views over the course of several hours. Following the conclusion of the slideshow, the photo frame stops at the latest image and continues updating at a rate of one image per minute. To further enhance the experience, we have included a button that senior residents can push if they are particularly drawn to a specific scene, allowing them to receive a real postcard from a slot featuring a corresponding image. The distribution of scenery-collectors to volunteers is a crucial aspect of the system as it enables the real-time sharing of local scenery with senior residents. The gallerylike interactive installation allows the residents to experience the scenery photos in a tangible and interactive way, which enhances their engagement with the content. A tangible user interface (TUI) is a user interface in which a person interacts with digital information through the physical environment [34]. The purpose of a tangible interface is to empower collaboration, learning, and design by giving physical forms to digital information, thus taking advantage of the human ability to grasp and manipulate physical objects and materials. By printing the photos into postcards, the system also facilitates communication between senior residents and the outside world. The provision of topics for communication could potentially reduce social isolation and loneliness among the residents. Figure 3 illustrates the practical application of the Interactive Gallery system, which can be replicated in other nursing homes or care facilities to improve the well-being of senior residents. Scenery-collectors for view sharing. The proposed system consists of specially designed camera kits, referred to as scenery-collectors, which are based on the Raspberry Pi platform (Figure 3). The camera kit has a robust brick-like cuboid appearance that is easy to position, and its shell is made of transparent acrylic covered with cement, making it waterproof and unobtrusive in outdoor environments. These kits are distributed to citizens in local communities who can place them wherever they choose to share scenic views with nursing home residents. Gallery-like installation used by senior residents. The gallery-like installation used by senior residents is placed in the nursing home and comprises three units, each equipped with one high-definition monitor, framed and decorated with a large round button below (Figure 4). The installation displays scenery photos, including location names transferred from the scenery-collectors, which can be printed as postcards simply by pushing the button. The postcards can be either exchanged among senior residents or sent to the citizens who shared the view. The study was conducted at a Dutch nursing home, where the Interactive Gallery system was deployed in the public space. A set of scenery-collectors were distributed to volunteers in the local community to share real-time scenery photos with the senior residents. The deployment of the system lasted for two weeks and involved two research methods: direct behavior observation and semi-structured interview. --- • Direct behavior observation was utilized to examine the potential behavior changes that the Interactive Gallery system could bring to the senior residents in the public space. Due to privacy requirements, video recording was not permitted on the research site. As a result, the researchers adopted direct behavior observation to capture the social interaction level of the senior residents. This method provided insight into when, where, how often a behavior occurred, and how long it lasted [35]. --- • Semi-structured interviews were conducted with the senior residents to understand their reflections on the use of the Interactive Gallery. The interviews were conducted in their own environments after the deployment, and the data analysis was deemed more realistic than laboratory data [36]. A pre-determined set of open-ended questions were used to prompt further discussion [37]. The analysis of the interview data provided valuable insight into the senior residents' experiences with the system and their suggestions for future improvements. --- Procedure The observation was carried out by two trained observers, namely the first and the third authors, with the aim of studying the behavior changes in senior residents due to the deployment of the Interactive Gallery system. The field study consisted of three rounds of observation, namely informal observation, baseline observation, and intervention observation, as illustrated in Figure 4. The observers used a standardized protocol to record the participants' behaviors. --- Research Site The average age of the senior residents in our selected nursing home was approximately 85 years old. Approximately twenty residents who suffered from dementia were not included in the study. The public space on the ground floor of the nursing home included a meeting area, a library, and a restaurant/café, which were open to all. Among these areas, the restaurant/café and meeting space were the most frequently occupied by senior residents. As such, we selected these two locations as the research sites, as shown in Figure 5. --- Data Collection of Observation Prior to the formal baseline observation, a short-term informal observation was conducted for one week to become acquainted with the observation area. During this period, observers recorded and summarized the behaviors of senior residents in the area (Table 1). --- Location The senior participants in the public space could be found in the canteen (Area B), meeting space (Area C), or outside. To better record each participant's location changes, each desk within the canteen and meeting space was numbered --- Behavior --- Social behaviors Chatting, billiards, order (food), and playing cards --- Passive behaviors Reading, eating, daze, drinking, watching billiards, watching cards, and smoking (outside) --- Other Infrequent or difficult to designate, includes watching laptops, sending letters, taking pictures, etc. --- Time duration For each participant, their time spent in each behavior and location was observed and recorded. Duration was recorded in increments of one minute, and less than one minute was recorded as one minute We recruited a sample of eleven senior residents who frequently appeared in the public area. According to our one-week-long informal observation, we observed a total of around ninety senior residents that appeared in the public space, however, not all of these residents could be eligible to be participants. The reason for this was that most of these senior residents appeared intermittently, which means they appeared in public spaces only a few times a week, and stayed in the public spaces for short durations. However, as we focused exclusively on the elderly peoples' behaviors in public areas, therefore, the behavior data for those who did not appear in the public areas was extremely limited, and they were less suitable candidates for the study. Based on the above, our selection criteria were threefold: First, participants had to frequently appear in public areas, these senior participants appeared in public space almost every day. This high frequency of public space attendance facilitated their increased involvement in our field study, allowing us to capture a more comprehensive understanding of their behaviors in these environments. Second, participants had to consent to anonymous observation. Third, we excluded individuals with significant physical impairments as an additional criterion. Additionally, we treated participants' gender, educational background, and previous occupations as random variables in our study. Observation of senior residents in Areas B and C (the canteen and meeting area, respectively) was conducted by two observers for a duration of two hours each day, excluding activity days in the nursing home. To minimize observer bias, unobtrusive observation techniques were employed, with observers sitting inconspicuously in corners of the public areas and refraining from interacting with the senior residents. The aim of this method was to make the act of observation and recording unnoticeable to those being observed [38]. Senior residents were free to engage in their usual activities without any disruption. --- Descriptive Data of Observation The descriptive data in our study were obtained from a previous work [5], and were based on the model presented by Gottesman [39]. We divided the activities of senior residents into two categories, social Behaviors and passive Behaviors. We recorded various information about the senior residents, including their locations, behaviors, and time duration of their activities. A detailed description of the observation data and the observation checklist can be found in Figure 6. --- Baseline Observation Results Before the deployment of the Interactive Gallery, we conducted a baseline observation to measure the social behavior of eleven participants in the public space. This initial observation was carried out for two weeks, in order to establish a starting point for the study. As shown in Table A1 we calculated the percentage of time that each participant spent engaging in different behaviors and locations. This was done by adding the number of observations in each category and dividing it by the total number of observations for that participant. The results show that the participants spent most of their time engaging in passive activities in the canteen (Area B), such as reading, eating and drinking, and watching billiards. This finding is consistent with previous research [5], which indicates that passive activities are a common form of social behavior among elderly individuals. --- Design Intervention 4.3.1. Deployment of the Interactive Gallery Our observations indicated that a majority of senior residents frequented the dining hall in Area C daily, walking along the corridor to participate in public activities. To cater to this foot traffic, we strategically deployed an Interactive Gallery composed of three units along the corridor wall in this area, as illustrated in Figure 7. The deployment was designed to last for two weeks and required no intervention from the researchers. The senior residents were free to interact with the units at their leisure. The scenery-collectors used in our study were simple to operate, making our recruitment process for volunteers straightforward. We recruited three individuals from our university and local communities who expressed an interest in sharing scenic photographs with senior residents. For instance, one of our volunteers shared a photograph of an alpaca from the zoo, accompanied by a message that read, "I am sharing this picture of an alpaca because I find it adorable, cuddly, and gentle by nature. I hope that this image will bring you joy and brighten up your day". Our field study procedure involved the following steps: firstly, we provided a detailed introduction to the project and distributed scenery-collectors to the volunteers. Next, we advised the volunteers to place the scenery-collectors in natural landscapes with minimal pedestrian traffic to avoid violating the privacy of others in outdoor photographs. The chosen locations included a landmark in the city center, a lake in the northern city, a river that ran through the university, a zoo in the northern city, a building in the southern city, and the city stadium. Lastly, each volunteer was requested to include a brief introduction to the scenery and their contact information (name and address) while sharing the photographs with the senior residents. This added a personal touch to the scenic photographs and facilitated the establishment of a connection between the volunteers and the older adults. --- Intervention Observation The researchers conducted an intervention observation lasting two weeks after the Interactive Gallery was installed. The observation data collected during the intervention phase was compared with the baseline data to determine if there was a significant difference in the eleven senior participants' social interaction behaviors. During the intervention observation, the researchers included an additional behavior category, "Interacting with Interactive Gallery," which comprised three subcategories: watching Interactive Gallery, printing postcards, and sharing postcards. These behaviors were recorded alongside the behavior categories of the baseline observation. Table A1 presents the results of the intervention observation. By analyzing the data, the researchers were able to gain insights into the senior residents' interactions with the Interactive Gallery, as well as its impact on their social behavior. --- Comparing Baseline and Intervention Observations of Senior Participants' Behavior Our hypothesis is that the deployment of the Interactive Gallery leads to an increase in social interaction among participants. More specifically, it is believed that the eleven senior participants allocated a greater amount of time to conversing with one another and also increased their time spent in the meeting area where the Interactive Gallery was situated. Therefore, we conducted a paired-sample t-test analysis to compare two sets of data: baseline observation and intervention observation. The first comparison focused on the percentage of time spent chatting, and the results show a significant difference between the two conditions. Specifically, the mean percentage of time spent chatting increased from 26.89% (SD = 21.61) in the baseline condition to 28.00% (SD = 22.47) in the intervention condition, t(10) = -2.25, p = 0.048. These findings suggest that the installation of the Interactive Gallery had a positive impact on social behavior, as the eleven senior participants spent more time chatting with each other. The second comparison focused on the percentage of time spent in the meeting area (Area C) in both conditions. The results show a significant difference between the baseline and intervention conditions, with a significant increase in the percentage of time spent in the meeting area (where the Interactive Gallery was located) in the intervention condition. Specifically, the mean percentage of time spent in the meeting area increased from 0.03% (SD = 0.05) in the baseline condition to 0.33% (SD = 0.44) in the intervention condition, t(10) = -2.34, p = 0.042. These results suggest that the installation of the Interactive Gallery in the meeting area was effective in attracting more residents to this space, thereby promoting social interaction among them (Table 2). The results suggest that the eleven senior participants spent a significantly higher percentage of time chatting and in the meeting space (Area C) during the intervention when Interactive Gallery was placed. This finding, along with the interview results, suggests that the Interactive Gallery had a positive impact on the social interaction of senior residents in the public space of the nursing home. --- Senior Residents' Usage Frequency of Interactive Gallery In addition to the observations and interviews, we also tracked the number of times senior residents printed postcards using the Interactive Gallery each day from 13:00 to 15:00. The results are presented in Figure 8, showing a peak on the fourth day and a gradual decline thereafter. This trend suggests that the initial attractiveness of the Interactive Gallery increased but did not sustain over time. This finding highlights the need to consider strategies for maintaining and enhancing the engagement of senior residents with the Interactive Gallery. --- Post-Study Interview In this study, we conducted interviews with thirteen residents to gain insight into their use of the Interactive Gallery. The participants were selected through purposive sampling and their demographic information is presented in Table 3. None of the senior residents reported any significant physical impairments that may have affected their ability to use the Interactive Gallery. The participants' educational backgrounds and previous occupations were considered as random variables to ensure a diverse sample. The interviews were designed to explore six key aspects of the participants' experiences with the Interactive Gallery (Table 3). The interviews were conducted in a semi-structured format to allow for flexibility and in-depth exploration of each aspect. The questions were open-ended, and participants were encouraged to elaborate on their responses. The data collected from the interviews were analyzed using thematic analysis to identify common themes and patterns across participants' responses. This method allowed us to gain a rich understanding of the participants' experiences with the Interactive Gallery and provided valuable insights for the development of future interactive installations. --- Post-Study Interview Procedure The purpose of the interviews was explained to the older adults The older adults were then asked to sign the consent form A video was shown to introduce the Interactive Gallery Each interview lasted 30 min and was conducted with care to collect comprehensive data --- Semi-structured interview topics --- Overall Overall impression and evaluation of Interactive Gallery The Interview results are as follows. --- Overall Impression The vast majority of interviewees provided positive feedback regarding their experience with the Interactive Gallery. They found it enjoyable and captivating, as one participant noted, "I was able to see places that I cannot physically access. It's a great experience". Many expressed their fondness for the scenery pictures displayed within the Interactive Gallery. The reasons for their appreciation can be summarized as follows: Firstly, due to their physical limitations, many of the participants were unable to leave the nursing home independently. As one interviewee stated, "There are two things that greatly impact my life: limited mobility and poor eyesight". (P7, M). Another participant shared, "I used to go to the city center often, but now it's challenging for me. As I am in the early stages of Parkinson's disease, my daughter does not feel comfortable allowing me to go alone" (P2, F). The Interactive Gallery provided them with a means of experiencing the beauty of the outside world from the comfort of their home. Secondly, the scenery photos stirred up fond memories for the residents. One interviewee, who was a former photographer, showed great interest, stating, "I find all of the photos quite lovely. Despite my love of photography and my possession of a small camera, my limited mobility now prevents me from taking pictures outdoors". (P5, F). Lastly, the residents were curious about the changes occurring outside of the nursing home, as one participant explained, "Having traveled extensively when I was younger, I am curious to see how the places have changed since then". The Interactive Gallery provided them with an opportunity to stay connected with the world outside of the nursing home. Overall, the positive feedback from the interviewees suggests that the Interactive Gallery has the potential to positively impact the well-being of senior residents in the nursing home. --- Interacting with Interactive Gallery Frequency of watching the Interactive Gallery. The frequency of watching the Interactive Gallery among senior residents varied, depending largely on their daily routines. Three interviewees reported watching it almost every day, as they regularly went downstairs. One interviewee explained her daily routine: "Each day, I typically make my way downstairs to the public space to enjoy a cup of coffee and check for any new updates to the Interactive Gallery display"-(P8, M). Seven interviewees used the Interactive Gallery approximately every other day. Three others watched it occasionally, either due to infrequent visits to the public space or a gradual loss of interest. As one interviewee stated, "The scenery images stay the same in most cases. I feel it is unattractive afterwards". (P2, F). Scenery photos of Interactive Gallery. Recognition of scenery photos in the Interactive Gallery varied among the interviewees. Most of them, except for two, recognized that the photos were of local city scenery due to their familiarity with the area. For instance, one interviewee stated, "I believe the first picture is of the northern city lake. I used to visit the place frequently, but that was a long time ago when I was with my children". (P11, F). On the other hand, the two interviewees who could not recognize the scenery were non-natives who originally came from other cities. As one of them explained, "I am not local, I spent only a little time here, I am not familiar with this city" (P9, F). However, some of the scenery photos were difficult to recognize due to significant changes in the scenery over the years. For instance, one interviewee commented on the photo of the zoo, "When I used to live there, the area was barren with nothing but sand. However, since then, a zoo has been established, and it is home to donkeys and giraffes". (P5, F). Therefore, the titles attached to the photos were essential clues for the interviewees to recognize them. As one of them pointed out, "I used to frequent that place often, but it has been a long time. The changes were so significant that I didn't recognize it at first, until I finally noticed the address". (P12, F). Most interviewees did not realize that the scenery photos were real-time since the picture changes were subtle. Hence, changing the frequency of the photos could increase their attractiveness. Additionally, it was noted that the frequency in which they watched the Interactive Gallery varied and depended on the interviewees' daily routines. While three interviewees watched it almost every day because they went downstairs frequently, seven watched it every other day, and three watched it occasionally due to lessening attractiveness. Understanding user interactions with the Interactive Gallery. During the study, most of the interviewees interacted with the Interactive Gallery by pressing the button to print postcards. However, there were two interviewees who did not press the button. One interviewee mentioned that the Interactive Gallery was too novel and unfamiliar to her: "I have never seen this kind of installation before, so I am not sure how to operate it". (P12, F) The other interviewee felt anxious when using it: "To be honest, I was hesitant to press the button, so I asked my son to print the postcards for me". (P4, M). Among the interviewees who did use the Interactive Gallery, some attended the introduction day and understood how to operate it. However, others simply followed suit and pressed the button without understanding its purpose. One interviewee stated, "Initially, I was unsure about the purpose of the button and simply followed suit when I saw others pressing it". (P10, F) In addition, three interviewees expressed that they did not press the button very often because they found the printing process time-consuming. Overall, the findings suggest that familiarity with the Interactive Gallery plays a crucial role in user interactions. While some users may require an introduction or a demonstration to understand how to operate the installation, others may feel too anxious or unfamiliar with the technology. Additionally, the time required for printing can also affect user behavior. --- Content Preferences Figure 9 illustrates the interviewees' content preferences. Firstly, the majority preferred hometown scenes, especially those where they were born or spent their childhood, as these places were full of cherished memories that were now difficult to access. This reflects their nostalgic feelings, as one interviewee expressed: "A picture of a mill would be particularly interesting to me. It would bring back memories of when I used to play with a doll carriage and run around the mill when I was twelve". (P11, F). Other interviewees echoed similar sentiments: "Seeing those pictures again brought back memories from the past. I miss those times, but now that I can't drive anymore, it's difficult to go back there. The pictures takes me back to the moments when I used to go to the park with my kids". (P10, F). Another shared, "My childhood home was situated behind a garden and in front of cornfields. I have fond memories of spending time outdoors there, and I miss being able to do so."-(P9, F). One interviewee from Indonesia had an album of hometown photos and shared, "Whenever I miss my hometown, I find solace in looking through old photo albums. I think about my hometown every day and long to visit again someday. It is my hope that I will have the chance to return and see my hometown once more."-(P12, F). Another interviewee stated, "I have a deep interest in preserving the memories of my hometown and have compiled three picture books of the area. I enjoy sharing them with others". Secondly, the interviewees preferred dynamic pictures over static ones. Scenery photos containing animals were particularly popular, as they made the photos appear more vibrant. As one interviewee explained, "I found the ones with zoo animals particularly captivating because they feature dynamic and ever-changing subjects. But the static scenery pictures don't offer the same level of excitement".-P4, M. Additionally, animal photos reminded them of their own pets. Some interviewees preferred photos from a kindergarten, which could bring back fond memories of their childhood: "Watching pictures of children playing brings back fond memories of my own childhood. It's as if I'm able to relive those moments of youthful innocence".-P6, F. --- Sharing of Interactive Gallery and Postcards Talking about Interactive Gallery with others. When asked about whether they shared the Interactive Gallery with others, the majority of interviewees mentioned discussing the scenery photos with other fellow residents, caregivers, and family members. For example, one interviewee said, "The cute alpacas reminded me of my cat, and I talked about it with someone next to me". Another interviewee shared, "I told my friend here that there is TV-like stuff downstairs, and I also mentioned it to my caregiver". Additionally, one interviewee shared the installation with their family members, stating, "I shared the installation with my son and we interacted with it together". These responses indicate that the Interactive Gallery not only has the potential to provide entertainment for residents, but also to facilitate social interaction and conversation between residents and their peers, caregivers, and family members. Use of postcards. Although all interviewees expressed liking for the postcards, they surprised us by revealing that they preferred to share the postcards with their family members face-to-face rather than writing on them. This preference was largely due to physical limitations, as one participant shared, "We grow older and have difficulties writing. In my case, I have sight loss, and my hands tremble". (P7, M). Two interviewees who did send postcard

The number of older adults residing in nursing institutions is increasing, and many of them experience social isolation. The social interaction of older adults constitutes a complex system that involves multiple stakeholders, including fellow residents, caregivers, members of the local community, etc. This paper proposes an Interactive Gallery, comprising a cluster of scenery collectors and an interactive installation resembling a gallery. It aims to promote social interaction among nursing home residents and members of the local community, as well as between senior residents within the nursing home. We conducted a field study that employed behavior observation and semi-structured interviews. Our findings show that the Interactive Gallery had a positive impact on the social interaction of senior participants, and it also stimulated their interest in sharing their experiences with individuals outside of the nursing home. The implications of our field study are significant. We highlight the social interaction system and behavioral characteristics of senior residents, strategies for enhancing social interaction within the nursing home, and strategies for promoting social interaction between senior residents and members of the local community. The Interactive Gallery presents a novel approach to addressing the issue of social isolation among senior residents in nursing homes. Our field study findings demonstrate its potential to improve the quality of life of seniors by promoting social interaction and engagement.

mentioned discussing the scenery photos with other fellow residents, caregivers, and family members. For example, one interviewee said, "The cute alpacas reminded me of my cat, and I talked about it with someone next to me". Another interviewee shared, "I told my friend here that there is TV-like stuff downstairs, and I also mentioned it to my caregiver". Additionally, one interviewee shared the installation with their family members, stating, "I shared the installation with my son and we interacted with it together". These responses indicate that the Interactive Gallery not only has the potential to provide entertainment for residents, but also to facilitate social interaction and conversation between residents and their peers, caregivers, and family members. Use of postcards. Although all interviewees expressed liking for the postcards, they surprised us by revealing that they preferred to share the postcards with their family members face-to-face rather than writing on them. This preference was largely due to physical limitations, as one participant shared, "We grow older and have difficulties writing. In my case, I have sight loss, and my hands tremble". (P7, M). Two interviewees who did send postcards included memories related to the scenery photos and expressed curiosity about the sharers, with one participant stating, "I am interested in learning more about the individuals who share their scenic experiences with us" (P4, M). For those who did not send postcards, concerns centered around privacy, time constraints, and lack of confidence. Some participants believed that their memories were too personal to share with volunteers, while others found postcard writing to be timeconsuming and tedious compared with other communication tools, such as phones. As one interviewee stated, "I used to write cards, especially around Christmas. However, nowadays, cards are becoming less common and people tend to use phones instead. As a result, I don't find many occasions to send letters". (P4, M). Another participant expressed concern that volunteers might not be interested in her memories and expressed a lack of confidence, stating, "I think I might not be smart enough, but if it's easy, I would still like to try volunteering". (P3, F). Fostering the senior residents' willingness to share their life with others. The interviewees showed a strong inclination to share something interesting with the volunteers despite not sending postcards back to them. One of the interviewees suggested: "It would be nice if we could take a camera and share our perspective, such as the children playing football outside". Additionally, they expressed gratitude towards the sharers and emphasized the importance of acknowledging the efforts of the volunteers who have done so much for them. One interviewee mentioned, "The photos are beautiful, and it's great that they share their views. We should thank the volunteers who have done so much for us". By acknowledging and appreciating the efforts of the volunteers, senior residents could become more motivated to engage with them and share their experiences. --- Suggestions for Improvements and Usability In terms of suggestions for improvements, the first issue is to enhance the immersion of the Interactive Gallery by including other sensory channels. As one participant noted, "I can hardly feel real nature, no sound, no smell, no wind".-P3, M. Another participant suggested that sound could be added to the installation to enhance the experience: "I believe the installation aims to bring nature closer to those who cannot physically access it. However, I think adding sound could enhance the immersive experience". (P4, M). Secondly, the subsequent iterations of the Interactive Gallery should take into consideration the individual needs of its users. As one participant stated, "Everyone has different views. It is not easy to make it good for all people". (P3, F). The third suggestion is to ensure the sustainability of the Interactive Gallery. One interviewee expressed concerns about its long-term usage, saying that it may not be attractive for repeat visits. "At first, we found the installation nice, but if people have seen it once, they may not feel the need to visit again." (P7, M). Fourth, it was suggested that additional functionality could be added to it beyond displaying scenery pictures. For example, one participant proposed: "In addition to displaying scenery pictures, the screen could also show the latest news to provide more diverse content". (P7, M). Regarding usability, while the printing function was stable, it was noted that it was time-consuming, taking nearly two minutes. This should be improved to provide a more efficient and faster user experience. Overall, these suggestions can improve the user experience and the sustainability of the Interactive Gallery, while also considering the individual needs and preferences of its users. --- Discussion --- Senior Residents' Social Interaction System, Behaviour Characteristics, and Technological Mastery Diversity Understanding the social interaction system and behavior characteristics of senior residents is essential for designing effective interventions to improve their social wellbeing. Future research can explore additional factors that influence social interaction, such as the role of technology in facilitating connections between senior residents and their social circles. --- Social Interaction System and Behavior Characteristics of Senior Residents The social interaction of senior residents can be classified into two types based on their social circle: internal social interaction and external social interaction. Internal social interaction refers to connections with fellow residents and caregivers within the nursing home, while external social interaction involves connections with family members, friends, and people from the local community. Senior residents' social interaction can also be categorized by the activities they engage in, which can be formal or informal. Formal social activities involve structured events, such as concerts, bingo, and billiards, while informal social activities are more casual, such as chatting, exchanging greetings, and social encounters. Understanding the social behavior patterns of senior residents in these different systems is important for developing strategies to improve their social well-being (Table 4). --- The senior residents' social interaction system Formal social activity (concert, bingo, billiards, etc.) Most did not actively participate in group activities, and the activities were not effective in promoting meaningful social interactions among them --- Informal social activity (chat, greetings, etc.) They preferred to stay in their rooms instead of spending time in the public spaces of the nursing home The social interaction of senior residents is affected not only by their social system but also by their behavioral characteristics. Our observation and interviews revealed that the senior residents tend to maintain a stable daily routine and are relatively resistant to change. Their usage of the Interactive Gallery, a social media technology designed to promote social interaction, heavily depended on their daily routines. We also found that the behavior of some senior residents was influenced by the behavior of their peers. Some senior residents used the Interactive Gallery because they saw others doing so, and their behavior encouraged others to interact with it. This phenomenon can be viewed through the lens of social proof theory, which suggests that individuals determine what is suitable, appropriate, or correct behavior based on the extent to which it is demonstrated by others [18]. --- Internal Social Interaction To fully leverage this behavioral characteristic, we propose that a small group of relatively active senior residents could be initially motivated to participate in social activities. Their behavior could then influence their fellow residents and further motivate them to participate. Thus, understanding the behavioral characteristics of senior residents can help us design interventions that are more effective in promoting social interaction. --- Facilitating Social Interaction among Senior Residents within Nursing Home --- Balancing Publicity and Individuality in a Senior Community Installation The Interactive Gallery installation in the senior community setting demonstrated a unique combination of public and individual attributes. In terms of publicity, it was placed in a communal area and used by multiple residents, with the primary goal of fostering communication and social interaction among the senior residents. Observations revealed that the Interactive Gallery had a positive impact on the eleven senior participants' social behavior: the duration of their chatting activities and the time spent at the installation site both showed a remarkable increase. Despite its public nature, the Interactive Gallery also had individualistic features that provided a sense of ownership and personalization for each older adult. Specifically, each resident was able to choose what they wanted to print on their postcards, which could be seen as a process of "privatization". The personalized postcards allowed the senior residents to express their individual preferences and connect with the installation on a more personal level. Moreover, the postcards became a conversational piece that they could share with their family and friends during visits. This combination of publicity and individuality in the Interactive Gallery allowed it to enhance the social connections among the senior residents while also meeting their individual preferences. The personalized postcards provided a sense of ownership and increased engagement with the installation, further reinforcing its impact on social behavior within the community. --- Design Should Be Integrated into Their Daily Routines and Promote Their Existing Habits Designing interfaces for older adults to support their social interaction requires careful consideration of their life habits, routines, and preferences. To promote natural and non-intrusive social behaviors, the design should be integrated into their daily routines and promote their existing habits. For instance, the Interactive Gallery design concept was inspired by the habit of older adults to watch the outside world through a window. Therefore, the Interactive Gallery's watching function aligns with their current habits, making it more accessible to them. Additionally, older adults' daily routines are relatively stable, and therefore, the Interactive Gallery should be placed within their routine, especially where most social behaviors occur. Rather than imposing socialization activities on older adults, promoting their social behaviors should be natural and non-intrusive. In the case of the Interactive Gallery, their memories are evoked by the scenery photos, which trigger conversations among them, making social behavior natural and non-intrusive. Integrating social communication into the interface can further promote social interaction among older adults, by allowing them to communicate with each other seamlessly. Thus, the interfaces should be tailored to the needs and preferences of older adults, taking into account their life habits and routines, to promote natural and non-intrusive social behaviors. According to the socioemotional selectivity theory, individuals' social goals tend to shift towards the strengthening of existing emotional relationships in late life, rather than in the pursuit of new social partners [15]. However, our study revealed that senior residents have the potential and motivation to connect with people outside of their family members. They appreciated the efforts of the volunteers and expressed a willingness to share their lives with others. Our project served as a catalyst for their communication desires, making them aware of the necessity of social communication beyond the nursing home, and sparking initiatives for both sides. While they possessed the desire to share their lives with others, they lacked the appropriate tools and confidence to initiate communication. To address this, our project aimed to provide them with tangible interfaces that integrate social communication and can be easily incorporated into their daily routines. By designing tools that are intuitive and non-intrusive, we hope to encourage and guide senior residents in connecting with people outside of the nursing home. In future iterations, it would be beneficial to facilitate communication between residents and scenery-collectors to enable suggestions for photographic subjects. This would not only enhance the overall excitement and engagement with the installation but also foster a greater sense of community and interaction. Therefore, encouraging and guiding senior residents to connect with people outside of the nursing home requires a comprehensive approach that takes into account seniors' needs, preferences, and unique challenges. By designing intuitive and non-intrusive communication tools and offering guidance and support, we can foster social connections that enhance seniors' wellbeing and reduce feelings of isolation. --- Separating Tangibility from Communication The topic of separating tangibility from communication is of critical importance in understanding the effectiveness of traditional communication methods, especially for marginalized populations such as seniors. While tangible objects such as postcards can create a sense of ownership and engagement, they may not always be the most efficient or effective means of communication. The field study discussed in this content highlights the challenges that seniors face when using traditional means of communication, including high costs, difficulty in writing, and a strong desire to keep the postcards as personal mementos. To address these challenges, it is necessary to explore new and innovative communication methods that can facilitate quick and easy interaction. The adoption of direct audio recording and speech recognition technology can lower the use costs for senior residents and enable them to communicate more effectively with volunteers and other outside people. However, it is important to consider the unique needs and preferences of different populations when developing these communication methods. Moreover, to fully appreciate the effectiveness of different communication methods, it is crucial to consider the theoretical foundations of communication. Communication theories, such as the communication accommodation theory and the social penetration theory, provide a framework for understanding the processes and effects of communication. For instance, the communication accommodation theory [40] posits that individuals adjust their communication style to match that of their interlocutor, which can be particularly relevant for intergenerational communication. Meanwhile, the social penetration theory suggests that the depth and intimacy of relationships increase as individuals disclose personal information to each other, highlighting the importance of trust and open communication in fostering social connections [41]. Therefore, separating tangibility from communication requires a comprehensive understanding of communication theories and the unique needs of different populations. By embracing new technologies and taking into account the diverse preferences and contexts of different individuals, we can create more effective and inclusive communication methods that strengthen our social connections and reduce feelings of isolation. --- Senior Residents Act as Content Producers, and the Outside People Act as Memory Trigger Providers The interviews conducted in our study revealed that the senior residents' memories were triggered by familiar scenery photos, indicating the prevalence of nostalgia among them. This is in line with the place attachment theory, which suggests that places hold emotional significance for individuals beyond their geographical boundaries [42]. Given the strong emotional bonds between senior residents and the places they have visited, the Interactive Gallery system leverages scenery photos as memory triggers to promote social interaction. Our study demonstrates that senior residents have the potential to become content producers, as they possess treasured albums, souvenirs, and stories that they are eager to share with others. While the concept of senior residents as content producers is not novel [8], our study investigates how their contributions can help build new connections with outside people. Through the Interactive Gallery system, the volunteers share scenery photos with senior residents, which trigger their memories and encourage them to share their own stories. In this way, outside people can not only receive content, but also act as memory trigger providers. This collaborative process promotes mutual understanding and boosts the senior residents' self-esteem. Our study highlights the importance of leveraging senior residents as active participants in the content creation process, and the potential benefits of using memory triggers to facilitate social interaction. Future research could explore how other forms of media, such as audio and video, could be used to encourage content creation and promote social interaction among senior residents. The study also points to the potential of using other forms of media to encourage content creation and social interaction among senior residents. For example, audio and video recordings could be used to capture the senior residents' stories and experiences in a more immersive way. This would allow them to share their memories and experiences with others in a more engaging and interactive manner. Overall, the study highlights the importance of recognizing the value of senior residents as content producers and leveraging their contributions to facilitate social interaction. By providing opportunities for senior residents to share their memories and experiences with others, we can promote mutual understanding and strengthen the connections between generations. --- Conclusions, Limitation, and Future Work This paper explored the potential of facilitating social interaction between nursing home residents and citizens from local communities and social interaction among senior residents within nursing homes and between nursing home residents and citizens from local communities. The findings suggest that the Interactive Gallery positively impacted social interaction among the senior residents within the nursing home (internal social interaction) but did not achieve the anticipated results for external social interaction. Nonetheless, it did stimulate the senior residents' desire to share their lives with people outside the nursing home. This study also identifies the social interaction system and characteristics of senior residents and proposes strategies to facilitate social interaction among senior residents within the nursing home and between senior residents and local communities. Regarding limitations, our study has certain limitations that should be taken into account. Our sample size was limited to eleven senior participants who were selected based on specific criteria. It is important to note that our findings cannot be generalized to the entire population of senior residents who frequent public areas. Therefore, we emphasize the need to interpret our findings accurately within the context of this limited sample. Despite the small sample size, we conducted a relatively long-term field study over a period of one month and collected valuable data that provide significant insights into our research question. We acknowledge that the small sample size is a clear limitation of our study; however, we believe that our findings still contribute to the current understanding of the topic and provide a foundation for future research in this area. In terms of future work, this study suggests investigating the potential of senior residents as content producers and exploring other forms of memory triggers, such as albums and mementoes. For the Interactive Gallery itself, the study suggests incorporating real-time video and sound to improve the dynamic effects, and considering other sensory channels, such as smell and wind, to make the experience more authentic and immersive. It is beneficial for residents to verbally express the content they desire to include on the postcard instead of relying solely on written communication. Overall, this study contributes to the understanding of social interaction among senior residents and provides insights into the design of technology to facilitate social interaction in nursing homes. Future research can build on these findings to develop more effective interventions to address the social isolation and loneliness experienced by many older adults in nursing homes. --- Data Availability Statement: To protect the privacy of participants, the data will not be disclosed to the public. --- Conflicts of Interest: The authors declare no conflict of interest. --- Appendix A

The number of older adults residing in nursing institutions is increasing, and many of them experience social isolation. The social interaction of older adults constitutes a complex system that involves multiple stakeholders, including fellow residents, caregivers, members of the local community, etc. This paper proposes an Interactive Gallery, comprising a cluster of scenery collectors and an interactive installation resembling a gallery. It aims to promote social interaction among nursing home residents and members of the local community, as well as between senior residents within the nursing home. We conducted a field study that employed behavior observation and semi-structured interviews. Our findings show that the Interactive Gallery had a positive impact on the social interaction of senior participants, and it also stimulated their interest in sharing their experiences with individuals outside of the nursing home. The implications of our field study are significant. We highlight the social interaction system and behavioral characteristics of senior residents, strategies for enhancing social interaction within the nursing home, and strategies for promoting social interaction between senior residents and members of the local community. The Interactive Gallery presents a novel approach to addressing the issue of social isolation among senior residents in nursing homes. Our field study findings demonstrate its potential to improve the quality of life of seniors by promoting social interaction and engagement.

Background Ageing entails a complex process of adapting to a multitude of age-related physical, psychological, and social changes [1]. To effectively respond to this process, people create cognitive representations of ageing, i.e. ageing perceptions [2,3]. These perceptions can be both positive and negative (e.g. [2]), and they can influence older people's well-being [4,5], health [2,4,6], quality of life [7,8], loneliness [9], anxiety [10], and depressive symptoms [2,10,11]. Studies of ageing perceptions have investigated mainly individual differences therein in Western populations. We need an in-depth understanding of cross-cultural differences in ageing perceptions, given the evidence for such differences [12,13]. Turkish migrants who arrived initially in the late 1960s and the 1970s currently form one of the largest groups of ethnic minorities in Western Europe. They show one of the strongest ageing trends and have been identified as one of the most vulnerable elderly groups [14]. Although older Turkish people in Europe represent an important study population, research investigating cultural differences in ageing perceptions between them and Western European natives is scarce. --- Cross-cultural differences in ageing perceptions Previous research has suggested that ageing perceptions are not necessarily ethnically neutral [15]. A growing body of literature has provided evidence that these perceptions differ among cultures [16]. According to social representations theory [17,18], views on ageing within a given culture are a form of shared cultural representation. They constitute systems of ideas, values, and customs related to ageing that are treated by members of the society as if they were established reality. As such, they are a mixture of accurate depictions of agerelated changes and the views of older people in general [19], which are expected to differ between older Turkish people and natives in Western Europe [13]. For example, old age is considered to begin at a much earlier stage according to Turkish people than people from western European cultures [20,21]. Furthermore, studies suggest that no consensus has been reached as to what constitutes being 'old', and that cultural differences may explain differences in this and other ageing perceptions. Furthermore, Turkish migrants are likely to have fewer resources compared with elderly native Europeans, which may cause difficulties in coping with old age. They might thus be at greater risk of developing maladaptive ageing perceptions, as supported by a few studies showing that elderly Turkish migrants felt old at a younger age [22] and had more adverse evaluations of their own ageing [23,24]. The extent to which these differences in ageing perceptions can be attributed to differences in culture or resources, however, remains unclear. --- Resources Perceptions of ageing play a vital role in older people's ability to adjust adequately to loss and major life events (e.g. the loss or sudden severe illness of one's spouse) [25]. Negative perceptions of ageing (i.e. having a severe loss frame) negatively influence adjustment to such life events because they limit the ability to see opportunities for substitution. Negative ageing perceptions, in turn, are affected by the availability of resources; more resources are associated with the ready availability of more opportunities, which are likely to weaken the loss frame. Research has shown that a stable and varied repertoire of such resources allows people to achieve instrumental goals, which are important for the achievement of overall well-being and compensation for losses that are associated inextricably with ageing [25,26]. An important physical resource is health. The optimisation of opportunities for good health enables older people to take part in society actively and enjoy independent and highquality lives [27]. Financial security has been identified as important economic resource, given that adapting to the process of ageing is generally easier when a person is without financial worries. Financial means help older people to feel less negative and worried, enabling them to cope better with ageing and life in general. Money, for example, helps not only to secure one's basic physical needs (e.g. food, drink, warmth), but also to compensate for declines in physical health by enabling a person to acquire at-home formal assistance. Education is an important personal resource, as it provides intellectual skills to deal with the consequences of ageing and to maintain a positive frame of mind [28]. Researchers have underscored the importance of higher educational levels for the development of more favourable ageing perceptions [29,30]. Genderbased differences in opportunities and living conditions are also known to accompany people across the lifespan. Older women in Europe, for example, are at greater risk of poverty than are older men [27]. Gender is evident as a personal resource in the context of life expectancy and cause of death. Although women tend to have longer life expectancies than men, they do not necessarily enjoy more years of better health or quality of life. The gender difference in life expectancy is thus smaller when years lived with disability are taken into account, implying that many women live longer with chronic disease or disability. Furthermore, older women in Europe report more psychological symptoms and greater use of medical consultation and treatment than do older men [27]. These differences may also result in different ageing perceptions. Actual age is another important personal resource; with increasing age, the ageing experience is increasingly framed in terms of physical decline or social loss and less in terms of continuous growth [28]. Social resources, such as having a spouse, children, and close friends, are also key in old age [1,25]. Becoming old may not be perceived as alarming when a person can rely on others, for example, to help with groceries or provide a listening ear. In a study conducted among older Korean Americans, social resources were found to contribute to more positive views about ageing [31]. Depending on the availability of such resources, older persons are more or less likely to have positive (or negative) perceptions of ageing. Access to a variety of resources may reduce worries about the impact of ageing on one's life and promote more positive ageing perceptions [25,26,28]. More specifically, those with a variety of resources more often think of old age as a time of development and less often associate ageing with physical decline and dealing with losses [28]. However, being a migrant or a member of an ethnic minority group can have a large impact on the accumulation of the aforementioned resources [32]. For example, long-term employment in low-paid manual-labour positions may lead to deficits in physical and economic resources, as one accumulates less financial means and may experience work-related physical discomfort in old age. Furthermore, being poorly educated may, for instance, impede understanding of the right to and procedures of elderly care [32]. On the other hand, migrant or ethnic minority elders may have more social resources than do native elders, due to strong ethnic bonds and cultural norms or values that promote interdependency and discourage divorce [32]. The greater availability of these social resources may offset some of the negative effects of the lack of other resources, thereby encouraging a more positive or less negative evaluation of ageing. In summary, some evidence suggests that personal, physical, economic, and social resources, in addition to culture, can explain cross-cultural differences in ageing attitudes. However, findings are limited. As populations in nations across the world age [33], societal perceptions of older adults and people's expectations about their own ageing processes are expected to influence the wellbeing of older adults, as well as policy decisions at the societal level. A better understanding of cross-cultural differences in ageing perceptions is a crucial step forward in navigating these challenges [13]. More research on the cultural meanings and significance of resources in relation to ageing perceptions in different populations is needed. Our lack of knowledge about cultural factors related to ageing may lead to the failure to plan appropriately for healthy ageing. We thus undertook the present research to compare ageing perceptions of natives and migrants in Rotterdam. --- Purpose of the study This study quantitatively examined ethnic differences in ageing perceptions in a large sample of migrant Turkish and native Dutch elders residing in Rotterdam, and explored whether such differences could be attributed to culture or resources (personal, physical, economic and/ or social). Based on previous findings, we hypothesised that ageing perceptions would be more negative among Turkish than among native elders (hypothesis 1). Furthermore, we expected that health (physical resource), education (personal resource), and income (economic resource) would be worse (hypothesis 2), but that social resources would be stronger (hypothesis 3), among Turkish than among Dutch elders. --- Methods --- Data collection This study was based on combined data from two research projects focusing on the health and well-being of community-dwelling elderly people in Rotterdam, the Netherlands. Ethnic background was defined according to respondents' countries of birth (i.e. Turkish for those born in Turkey, native Dutch for those born in the Netherlands). The first dataset was gathered in February-April 2012 and contained data from 994 native Dutch elders aged 70-99 years [34]. The Rotterdam municipal register was used to randomly sample respondents, stratified by age group (70-74, 75-79, 80-84, and <unk>85 years) and neighbourhood. The number of participants per neighbourhood was weighted according to the population ratio of the district. Only one person per address was allowed to participate. Participants were sent questionnaires with written invitations and self-addressed envelopes. Two reminders were sent in cases of non-response. Of 2890 people approached, 297 were excluded because they did not meet the study criteria [i.e. were not community dwelling (n = 67) or native Dutch (n = 205)] or were unable to participate due to serious medical problems (i.e. dementia) or death (n = 25). Of the 2593 eligible respondents, 1075 returned filled-in questionnaires (41% response rate). Of these 1075 respondents a total of 994 were natives which we used for the current study. The second dataset contained data from 680 Turkish migrants aged 65-90 years, gathered in March-June 2015 and October 2015-February 2016. Data were not collected between June and October 2015 because many elderly Turkish individuals travel to Turkey during this period. All Turkish people aged <unk>65 years were identified using the Rotterdam municipal register and invited to participate (census data). More than one person per address was allowed to participate. In total, 2350 respondents were sent questionnaires with written invitations and self-addressed envelopes. In the first data collection period (March-June 2015), the invitation and questionnaire were made available in Dutch and Turkish. As we noticed that the majority of respondents could not read Dutch, only the Turkish versions of the questionnaire and invitation were sent to respondents in the second data collection period (October 2015-February 2016). In cases of non-response, single reminders were sent, followed by at-home visits by ethnically matched interviewers. In total, 680 Turkish respondents returned filled-in questionnaires (32% response rate). Of the non-respondents, 213 were ineligible due to serious medical issues or death (n = 102), change of address (n = 110), or non-Turkish ethnic background (n = 1). --- Measures --- Perceptions of ageing Ageing perceptions were measured using the 21-item Ageing Perceptions Questionnaire-Short (APQ-S) [2,4]. The APQ-S assesses ageing perceptions across seven (sub)dimensions, based on Leventhal's self-regulation model [35]: timeline, or a person's awareness of ageing, which can be constant (chronic/acute: e.g. "I always classify myself as old") or vary over time (cyclical: e.g. "I go through phases of feeling old"); consequence, or the believed impact of ageing on various life domains, which can be positive (e.g. "As I get older, I get wiser") or negative (e.g. "As I get older, I can take part in fewer activities"); perceived control over aspects related to ageing, which can be positive (e.g. "Whether I continue living my life to the full depends on me") or negative (e.g. "Slowing down with age is not something I can control"); and emotional representations, comprised of negative emotional reactions toward one's ageing (e.g. "I get depressed when I think about getting older"). Each dimension has three items, with responses ranging from 1 ('totally disagree') to 5 ('totally agree'). The items of the control negative dimension were reverse coded, so that higher scores indicated more perceived control. --- Physical resources Respondents were asked to rate their perceived current general health on a five-point scale ranging from 1 ('very bad') to 5 ('very good'). This measure is considered to be valid and robust; self-reported health has been proven to be a strong predictor of mortality, physical disability, and chronic illness [36][37][38]. Responses were dichotomised as indicating 'poor' (1, 2) and 'good' (3-5) health. --- Personal resources Respondents were asked to report the highest educational level completed in the Netherlands or abroad. For the Dutch sample, educational level was characterised using seven categories, ranging from 'less than 6 th grade of elementary school' to 'higher vocational school and university'. For the Turkish sample, two additional categories were used: 'no schooling' and an option to write down other, unlisted, forms of schooling. This variable was dichotomised as completion of elementary school or less (1) or more than elementary school (0). The questionnaire also solicited information on age and gender. --- Economic resources Income level was determined based on respondents' reported monthly household income, including social benefits, pensions, and alimony. Responses ranged from 1 ('less than €1000 a month') to 4 ('€3050 a month or more'). 'Do not know/do not want to tell' was included as a fifth category. Income level was dichotomised, with '€1350 or more' used as a reference category (below '€1350 in the Netherlands is considered to be below a minimum standard (Blommesteijn and Mallee 2009)). --- Social resources Respondents were asked to indicate whether they were married, divorced, widowed, single, or living together as one half of an unmarried couple. A dichotomous variable was created, with 'divorced, single, or widowed' serving as the reference category. --- Analyses Sample characteristics were examined for the total study sample and separately for Dutch and Turkish respondents. To explore hypotheses 1-3, the significance of differences in ageing perceptions and resources between the Turkish and Dutch samples was tested using independent-samples t tests and chi-squared tests. Bivariate associations were calculated to investigate the relationship between resources and ageing perceptions. Finally, stepwise regression analyses were performed for each dimension of ageing perception. Each regression analysis consisted of two steps. In the baseline model (model I), only ethnicity (culture) was regressed onto the ageing perception dimension. In the second model (model II), physical (health), personal (age, educational level, and gender), economic (income), and social (marital status) resources were entered into the equation. The results of regression analyses were compared between the Dutch and Turkish samples to investigate the relative contributions of the independent variables to all seven ageing perceptions. --- Results --- Sample characteristics The Turkish sample was significantly younger than the Dutch sample [72.89 (5.03) vs. 78.66 (6.07) years; p <unk> 0.001] and included fewer women (47.6% vs. 58.1%; p <unk> 0.001; Table 1). Large differences were found between Turkish and Dutch older people in educational level (low, 80.9% vs. 15.5%; p <unk> 0.001) and monthly household income (low, 84.7% vs. 34.2%; p <unk> 0.001). Furthermore, more Turkish than Dutch older people reported being in poor health (72.6% vs. 44.5%; p <unk> 0.001). Finally, fewer Turkish than Dutch elders were single, divorced, or widowed (29.1% vs. 57.4%; p <unk> 0.001). --- Ethnic differences in ageing perceptions As expected, significant differences in ageing perceptions were found between Turkish and Dutch respondents (Table 2). On average, Turkish respondents reported a more negative awareness of ageing, as shown by significantly higher mean chronic/acute and cyclical timeline scores. Furthermore, Turkish respondents felt significantly less in control of positive and negative aspects of the ageing process. Turkish elders also reported having significantly more negative emotional reactions to ageing, such as feeling down or depressed and angry. Finally, whereas Turkish elders believed more firmly that the ageing process would have negative impacts on their lives (e.g.'As I get older I can take part in fewer activities' and 'Getting older makes everything a lot harder'), they also experienced the positive consequences of ageing (e.g. 'With age comes wisdom, personal growth, and feeling more appreciative of things') more strongly than did native Dutch elders. Ageing perceptions: a matter of culture or resources? As ethnic differences were found across the whole range of ageing perceptions, further analyses were performed to assess whether these differences are explained by culture or resources. Several significant but small associations were found between age and ageing perceptions (Table 3). For example, older respondents reported more chronic perceptions of their ageing and believed that their ageing had more negative and less positive consequences. A significant gender difference was found only for the negative control dimension; on average, male participants felt more in control of negative aspects related to ageing (e.g. becoming less mobile) compared with older women. Ethnicity was associated significantly with all ageing perceptions (Table 3). In general, Turkish ethnicity was associated positively with the negative ageing dimensions (chronic and cyclical timeline, negative consequences, and emotional representations) and negatively with the positive ageing dimensions (feeling in control over positive and negative ageing effects). The only exception, in line with the findings presented in Table 2, is the observed positive relationship between Turkish background and the positive consequences of ageing (wisdom, personal growth, and feeling more appreciative). As expected, monthly household income and educational level were also associated significantly with all ageing perceptions. The directions of these associations were similar to those observed for Turkish ethnicity; low education and income levels were associated positively with the negative ageing dimensions and negatively with the positive dimensions, except for the positive consequences of ageing. Respondents with low incomes and educational levels were more aware and felt less in control of their ageing processes. Furthermore, they reported having more negative emotional responses toward their ageing. As with Turkish ethnicity, these respondents believed more strongly that ageing had negative impacts on several life domains, but also that ageing had positive effects. Poor health was associated with all ageing perceptions except the positive consequences of ageing. Respondents who were in poor health had more negative and less positive ageing perceptions compared with those in good health. --- Multivariate associations between resources and ageing perceptions To examine whether ethnicity remained a significant predictor of ageing perceptions when the previously mentioned resources were taken into account, stepwise regression analyses were performed. Culture was related significantly to ageing perceptions in models I and II (Table 4). The direction and strength of these relationships were similar to those of the bivariate associations. Health, age, and education were also found to be important predictors of ageing perceptions. Although low income level was associated with all seven dimensions of ageing perceptions in the bivariate analysis, it showed no significant association in the adjusted analysis. Overall, low educational level was related to more negative and less positive ageing perceptions (e.g. stronger awareness of ageing and weaker belief in its positive influence). The proportion of explained variance was remarkably low for most ageing perception dimensions. In the final model, this proportion was lower in the Turkish sample (Table 5) than in the Dutch sample (Table 6); it reached 25% only for the negative consequence dimension. In the Turkish sample, poor health and education were the most important resources for ageing perceptions. In the Dutch sample, poor health and age were the most important resources; education was not associated significantly with any ageing perception. For the timeline chronic dimension, women had more negative ageing perceptions than did men. While the explained variance among the Dutch is higher compared to the Turkish it is still relatively low, which indicated other resources may be more important when it comes to ageing perceptions among older people. --- Discussion Although ethnic differences in ageing perceptions may be rooted in socio-economic and health differences, cultural differences may also shape the ageing experience [22][23][24]. Furthermore, research clearly emphasized the benefits of considering multiple predictors simultaneously when it comes to investigating ageing perceptions [39]. The results of this study clearly reveal the importance of culture for all ageing perceptions among Turkish and Dutch elders residing in Rotterdam. This is in agreement with earlier studies [22][23][24]39]. We additionally found that age, health, and education were also important factors. For Turkish elders, health and education were the most important resources; for Dutch elders, age and health were most important in relation to ageing perceptions. The explained variance in the final model was low overall and was greater in the Dutch than in the Turkish sample, indicating that factors not included in the model explain ageing perceptions among Turkish elders. Thus, additional cultural factors, such as acculturation [40] and religion [41], should be taken into account in future studies. Ageing perceptions were generally more negative among Turkish than among Dutch elders, supporting hypothesis 1. Turkish elders indeed reported more negative awareness of ageing, felt less in control of their ageing processes, and had more negative emotional reactions to ageing. They, however, believed more firmly that their ageing processes would have both negative and positive effects on their lives. Culture seems to lead to certain expectations regarding growing older which also include personal growth and feeling more appreciative of life. Regarding all the other six ageing perceptions our hypothesis was confirmed; overall a Turkish background leads to more negative ageing perceptions and less positive ones compared to native Dutch elderly. The study results also support hypotheses 2 and 3. They clearly revealed poorer health and lower income and educational levels among Turkish than among Dutch elders (hypothesis 2), in agreement with previous findings [32]. In addition, many more Dutch than Turkish elders were single/widowed/divorced (hypothesis 3). Although the response rate was lower for Turkish than for Dutch elders in this study, it was comparable to those in other surveys conducted in this population [42]. We made a minimum of two contact attempts for the Turkish sample. To improve the response rate, this number should be increased to six contact attempts, which was not feasible in our study [22,42]. Another limitation of this study is its cross-sectional design, which hampered the ability to draw causal conclusions. Thirdly, this study was conducted only among Turkish and Dutch elders. It would be interesting to conduct a similar study among, for example, Moroccan elders in Rotterdam, as they are the second largest migrant group in the city. Finally, we merged two datasets, which limited our choice of variables representing the resources examined. Future research should use a more extensive set of variables to capture a wider repertoire of resources. --- Conclusion We can conclude that physical and personal resources, in addition to culture, are most important for positive ageing perceptions among native Dutch and migrant Turkish elders residing in Rotterdam, the Netherlands. These resources are health and education for Turkish elders, and health and age for Dutch elders. Overall, ethnicity remained a significant predictor of ageing perceptions when other resources were taken into account. However, a large part of the variation in ageing perceptions remained unaccounted for, especially among Turkish elders. Thus, we have solved only a small part of this puzzle. Research investigating additional cultural aspects, such as acculturation and religion, is needed. Such efforts may increase our understanding of ethnic differences in ageing perceptions. --- Availability of data and materials The datasets analyzed during the current study are available from the corresponding author on reasonable request. --- Personal resources Age Author's contributions Both authors contributed to the design of the study, performed the statistical data analyses, interpreted the findings, and drafted and refined the manuscript. Both authors read and approved the final manuscript. --- Ethics approval and consent to participate According to the CCMO, the current study did not fall within the scope of the Medical Research Involving Human Subjects Act and therefore did not have to undergo prior review by an accredited Medical Research and Ethics Committee or the CCMO [43]. All respondents were informed about the aims of the study and its anonymous and voluntary nature, before giving their consent to participate. The dataset used for comparison was approved by the ethics committee of the Erasmus University Centre of Rotterdam in June 2011 (reference number MEC-2011-197). Prior to giving their consent, all respondents were informed about the aims of the study and its anonymous and voluntary nature. --- Consent for publication Not applicable. --- Competing interests Jane Murray Cramm is an associate editor of BMC Geriatrics. Otherwise the authors declare that they have no competing interests. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: This study examined ethnic differences in ageing perceptions of migrant Turkish and native Dutch elders residing in Rotterdam, and explored whether such differences could be attributed to culture or resources (personal, physical, economic and/or social). Methods: This study was based on combined data from two research projects focusing on the health and wellbeing of community-dwelling elderly people in Rotterdam. The first dataset contained data from 994 native Dutch elders aged 70-99 years. The Rotterdam municipal register was used to randomly sample respondents, stratified by age group (70-74, 75-79, 80-84, and ≥85 years) and neighbourhood. Of the 2593 eligible respondents, 1075 returned filled-in questionnaires (41% response rate). Of these 1075 respondents a total of 994 were natives which is the sample we selected for the current study. The second dataset contained data from 680 Turkish migrants aged 65-90 years. All Turkish people aged ≥65 years were identified using the Rotterdam municipal register and invited to participate. In total, 680 Turkish respondents returned filled-in questionnaires (32% response rate; out of 2350). Ageing perceptions were measured using the 21-item Ageing Perceptions Questionnaire-Short (APQ-S). Respondents were additionally asked about their current general health, income, education, marital status, age and gender. Results: The results of this study clearly reveal the importance of culture for all ageing perceptions among Turkish and Dutch elders. We found that age, health, and education were also important factors. For Turkish elders, health and education were the most important resources; for Dutch elders, age and health were most important in relation to ageing perceptions. Ageing perceptions were generally more negative among Turkish than among Dutch elders. Turkish elders reported more negative awareness of ageing, felt less in control of their ageing processes, and had more negative emotional reactions to ageing. They also believed more firmly that their ageing processes would have both negative and positive effects on their lives. Results revealed poorer health, lower income and educational levels among Turkish than among Dutch elders. In addition, many more Dutch than Turkish elders were single/widowed/divorced. Conclusion: Physical and personal resources, in addition to culture, are most important for positive ageing perceptions among native Dutch and migrant Turkish elders residing in Rotterdam, the Netherlands. These resources are health and education for Turkish elders, and health and age for Dutch elders. Overall, ethnicity remained a significant predictor of ageing perceptions when other resources were taken into account.

INTRODUCTION The health crisis due to Covid-19 means a change in ways of life both at the individual level and in the socio-political organization of the world's territories, so that the different sectors, such as the public, social and private have adapted various strategies to ensure survival in the turbulent environment that has plagued the planet since November 2019. In this context, governments at the federal, state, and local levels have an important role to play in protecting the health of the population, reducing the impact and lead processes of adaptation to the new normality. Mexican municipal governments, specifically those of the Oaxacan province, present characteristics of institutional weakness (D<unk>az, 2011); however, they also highlight exceptional cases where the management of the health crisis has been effective, due to the actions and results obtained. The starting point of the study is the legal obligation that municipal governments have to ensure the well-being of their population, in the Oaxacan case, such attribution is established in the third article of the Municipal Organic Law for the State of Oaxaca, which mentions that: "The City Council's primary mission is to serve the population within the legal framework for peace, equality between men and women, justice and social development", so in a crisis situation such as the pandemic, the local government has the duty to serve its population, establishing various public policies that contribute to the preservation of life and health. In article 44, the law identifies that the city council should not, "Issue agreements or actions that contravene provisions of authorities in health matters that put the health of the population at risk." In the same way, article 61 identifies as a serious cause for the revocation of the mandate of a member of the City Council: "Disobeying sanitary measures or public health protocols, in the case of health emergencies, epidemics, or any health crisis." These legally established restrictions show the obligation that the Oaxacan municipal governments have regarding the provision of public health in their territories, these are the main reasons for dedicating an analysis to the actions that the The leadership of the Oaxacan municipalities against COVID-19: The case of Santa Catarina Cuixtla, México, an analysis from the crisis management approach Oaxacan city councils have conducted to face the pandemic. In this order of ideas, the research question is: What effective strategies have the municipalities of Oaxacan indigenous regulatory systems implemented to face the Covid-19 pandemic? In such a way that the objective that corresponds to the question is: to analyze the effective strategies that the municipalities of indigenous normative systems of the state of Oaxaca have implemented to face the Covid-19 pandemic, through the crisis management approach. The study municipality is Santa Catarina Cuixtla, a population of Zapotec origin, located in the Sierra Sur region of the state of Oaxaca, whose sociopolitical organization regimen is the indigenous regulatory systems. Indigenous normative systems include collective norms that have been present and have survived for several centuries. This regime has its own political institutions which are differentiated and inveterate, such as the community assembly, the system of community charges and the tequio (in-kind cooperation and work by members of a community for a common goal such as public works, clean-ups, and parties). These systems include internal rules or specific procedures for the renewal of mayors and councils, in accordance with the federal and state constitutions. Each municipality in the regime has its own regulatory systems. Indigenous peoples maintain their systems traditional normative in collective acts and of relevant importance for community life, such as the work commissions, the tequio and the election of local authorities. This system of social and political organization, commonly known as uses and customs, grants its own cultural identity and represents an example of direct participatory democracy in government through the community assembly (Carlsen, 2003). --- MATERIALS AND METHODS This is a study with a qualitative approach, with an exploratory-descriptive scope, since it is a new phenomenon and characterizes crisis management (Boin, Kuipers & Overdijk, 2013), according to the design, it is both documentary and field research, because secondary sources were used to obtain information. Data from the context of the pandemic and for theoretical support, later field work was carried out, through a semi-structured interview instrument (Annex 1) in this sense five councilors were interviewed. Due to the particularity of the analysis, the case study was used (Stake, 1991& Prats, 2005), whose key informants were the members of the city council of the aforementioned municipality; In relation to temporality, it is a cross-sectional analysis (Tamayo & Tamayo, 2007), carried out in the months of September -November of the year 2021. In the first part of the document, a general overview of the pandemic is presented, as an emerging phenomenon, which took the world by surprise, before which both governments and civil society and individuals have designed strategies to lessen the effects, in this In this sense, some figures are presented on the effects of the disease at the global, national and local levels. Then, the theoretical approach that supports the document is addressed, taking Boin et al. (2013) and Culebro (2021) as main references, who have studied public policies and government actions from the perspective of crisis management, a scenario where frames the current health contingency. The theoretical approach is complemented with social cohesion, an important characteristic of native peoples. Next, the analysis of the strategies and actions carried out by the local government of Santa Catarina Cuixtla is presented. The analysis follows the order of ten executive tasks for effective crisis management according to Boin et al. (2013) namely: early recognition, meaning making, critical decision making, coordination, coupling and uncoupling, meaning generation, communication, accountability, learning and resilience. This theoretical approach was taken up again, as it was considered integral and ad hoc to the problem addressed, clearly establishing the categories of analysis. The municipality was chosen because its confinement measures made national news, specifically on May 15, 2020 (Andrew, 2020) when, in the midst of the pandemic, residents of the municipality, backed by the municipal authority, did not allow entry to the town of a funeral procession from the country's capital, one of the points hardest hit by the pandemic. This situation attracted the attention of the outside towards the management of the crisis of the city council, when observing the organization of the citizens and the city council, to protect the resident population in the municipality. --- RESULTS AND DISCUSSION --- This The world facing an unknown enemy, the Covid-19 At the end of December 2019, the media around the world announced the outbreak of a hitherto unknown coronavirus. It was in the city of Wuhan, Hubei province, China where the first case of infection was presented on November 1st, 2019 (Guzmán, 2021); however, it was not until December 31 of that year, when the Municipal Commission of Wuhan The leadership of the Oaxacan municipalities against COVID-19: The case of Santa Catarina Cuixtla, México, an analysis from the crisis management approach Health and Sanitation reported 27 cases of atypical pneumonia to the World Health Organization (WHO), including seven severe cases (Sanitaria 2000(Sanitaria, 2020)). In this regard, the World Health Organization (WHO, 2020) reported that the "causative agent of this pneumonia was identified as a new virus of the Coronaviridae family" called SARS-CoV-2, whose associated clinical picture is called Covid-19. The cases spread relatively quickly through several countries, so on January 30, 2020, the WHO declared the Covid-19 outbreak a public health emergency at the international and later global level. The first case in the American continent was reported in the United States, on January 20, 2020. For the Latin American region, Brazil registered the first contagion on February 26, 2020, since then the coronavirus has spread to the 54 countries and territories of the continent, according to the Pan American Health Organization (PAHO, 2020). In Mexico, on February 27 of the same year, patient zero was detected, and the first death from the new coronavirus in the country was registered on March 18 (Guzmán, 2021). By April 30, the number of patients had reached 19,224 confirmed cases and 1,859 deaths, that is, a mortality rate of almost 10 per cent, a relatively high figure (Suárez, Suárez, Oros, and Ronquillo, 2020). On March 11, 2020, the day the World Health Organization had declared the health crisis a pandemic, 118,000 cases had been detected in the world, with 4,291 deaths in 114 countries. A year later, on March 11, 2021, according to The Coronavirus Resource Center of Johns Hopkins University, the number of positive cases in the world had reached 117,971,734, with 2,618,000 deaths, in 255 countries and territories on the planet (Johns Hopkins University, 2021). The scenario for Mexico unfolded as follows: on March 11, 2020, 11 cases, and 0 deaths were recorded, exactly one year later, according to official figures, 2,144,558 cases with 192,488 deaths had occurred (Milenio, 2021), a critical situation for both society and governments. By September 17, 2021, deaths were counted at 322,582, which placed the country in fourth place overall in number of deaths worldwide and sixth place in deaths per million cases (Penna, 2021), with these the data reflected the low competitiveness in the management of the pandemic by the federal government. University database (2021) on October 08, 2021, showed 236,940,594 cases worldwide, with 4,837,772 deaths, and 6,418,953,258 doses of vaccines administered. The outlook for the state of Oaxaca for the same day showed a total of 78,554 cases, and 5,266 deaths (Johns Hopkins University, 2021). Faced with the new health challenge, the country's health infrastructure was overwhelmed, because even with the experience of Influenza AH1N1 in 2009 "no important public policies were generated to safeguard health. At the time the country only has a national institute for diagnosis and epidemiological reference, but it is not enough" (Guzmán, 2021) according to these data, the Mexican health system has been inefficient in responding to the demands presented for the pandemic. Some of the measures established by the governments of the world to curb the contingency were perimeter closures, strict curfews, or reduction of social activity (RTVE, 2021). From this new challenge, the governments of the countries followed four models of epidemiological intervention to face the effects of the coronavirus, China implemented the orthodox model, based on the isolation of the detected cases from the rest of the population to prevent the spread of the virus, lower the contagion curve and reduce cases, it is a restrictive, authoritarian and total isolation model, where borders are closed and massive reviews are carried out, resorting to isolation by quarantine, this model was also used in other countries such as the United States United States, Canada, and Italy (Hernández, 2021). South Korea leaned towards the population technology model, based on mass testing and the use of QR technology to monitor epidemiological progress. Under this model, massive tests are constantly used in strategic and broad sectors of the population, basically using PCR and antigen tests. The natural history model of the disease, commonly known as herd immunity, was used in countries such as England and Sweden. This method consists of keeping all the activities of the population open, allowing the free movement of people, which which will allow contact between them to generate some immunity to SARS-CoV-2. The general surveillance model for complex phenomena, also called sentinel monitoring, was used in Mexico. It consists of creating a group of surveillance sites, where the behavior of the disease and the number of cases are monitored, generating information to determine a health traffic light that determines the restriction or opening of essential activities. Around 347 sentinel centers were created in the country, located mainly in hospitals and health centers. According to the intervention model, Hernández (2021) presents the following figures (table 1) on the behavior of COVID-19 in some countries, including Mexico: Perform It is possible to observe that, although China has the largest population in the world, thanks to the orthodox model it has registered a lower number of deaths than Mexico, which has followed the sentinel model, from this it is observed that the policy that the Mexican federal government has followed to manage the pandemic has been the least efficient, compared to the models followed by other countries, in relative terms with respect to the number of cases and deaths. The mortality rate in Mexico is also due to complications from chronic diseases present in the population. --- The context of indigenous regulatory systems in Oaxaca The state of Oaxaca, Mexico has a vast cultural and political diversity, in this multicultural context, the Rural Water and Sanitation Information System (SIASAR, 2021) mentions that around 19 ethnic groups coexist in the entity: Afro-Mexicans from the small coast from Oaxaca, Amuzgos, Chatinos, Chinantecs, Chocholtecos, Chontals from Oaxaca, Cuicatecs, Huaves, Ixcatecos, Mazatecs, Mestizo, Mixes, Mixtecs, Nahuas, Tacuates, Triquis, Tzotzils, Zapotecs, and Zoques. According to information from the Secretariat of Indigenous and Afro-American Peoples (SEPIA, 2016), 65.7 percent of the Oaxacan population recognizes itself as Indigenous, 32.2 per cent speaks an original language, and multiculturalism is present in the eight regions of the state: Ca<unk>ada, Costa, Isthmus, Mixteca, Papaloapan, Sierra Norte, Sierra Sur, and Central Valleys. Such ethnic and cultural diversity has resulted in a social and political framework that is reflected in particular forms of government at the local level. In the political context, the electoral legislation for the state of Oaxaca recognizes two types of political systems for the local sphere, the municipalities governed by indigenous normative systems and the municipalities under the regime of political parties. In the first article, third section of the Law of Electoral Institutions and Procedures of the State of Oaxaca, it is established: "The recognition, safeguarding and guarantee of the democratic institutions and practices of the municipalities that are electorally governed by their indigenous normative systems". According to the second article, section XXIX of this legal system, the indigenous regulatory systems are:...the set of principles, oral or written norms, practices, institutions, agreements and decisions that the peoples, municipalities, indigenous and Afro-Mexican communities recognize as valid and in force for the election or appointment of their authorities and representatives, the exercise of their own forms of government and the resolution of internal conflicts. The State Electoral and Citizen Participation Institute of Oaxaca (IEEPCO, 2021), contemplates the indigenous regulatory systems, commonly known as uses and customs, refer to the general principles, the oral or written norms that the municipalities and Indigenous communities recognize as valid to apply in their daily life. In other words, the indigenous normative systems represent the forms of social, political, even economic, and cultural organization of the native peoples of the state of Oaxaca, manifesting themselves in daily coexistence, social interaction, forms of political organization, popular election, and the municipal government. Each municipality that is governed by this regime has its own regulatory systems; however, common and distinctive factors such as the community assembly, the cargo system and tequio, among others, persist. The Indigenous peoples maintain their traditional normative systems in force in collective acts of relevant importance for community life, such as work commissions, the tequio, and the election of municipal authorities. This system of social and political organization, where uses and customs stand out, grants its own cultural identity, and represents an example of direct participatory democracy in government through the community assembly (Carlsen, 2003). Indigenous regulatory systems comprise collective norms present in Indigenous peoples for several centuries, which are not infallible, but have proven their flexibility, coherence and ability to coexist with the modern State (Ram<unk>rez, 2013). In this regime, their own, differentiated, and inveterate political institutions have been developed, which include internal rules or specific procedures for the renewal of their city councils, in accordance with the federal and state constitutions in the aspect of the rights of Indigenous peoples. The election of municipal councilors is carried out according to the norms that the community has established over time, such as participation in the community assembly, the designation of contenders at the time of the election, the way of registering the votes, and the criteria for nominations among other things. The leadership of the Oaxacan municipalities against COVID-19: The case of Santa Catarina Cuixtla, México, an analysis from the crisis management approach Another characteristic aspect of the uses and customs is that the general community assembly is recognized as the main consultation body for public affairs, and where recognized citizens participate in full use of their faculties. When there are problems at the community level, the city council summons the population to meet in assemblies, where issues that concern all the inhabitants are discussed together, in such a way that important decisions are made collectively, and in an ideal context legitimize the actions of the municipal government. In Oaxaca, the Indigenous community is particularly numerous, representing almost half of the state population (Recondo, 2006), which is not homogeneous, since it belongs to 19 of the 56 ethnolinguistic groups present in the country, observing different cultural characteristics. The diversity of municipalities and the number of human settlements is mainly due to historical reasons related to the defense of the towns, their municipal and territorial categories, as well as cultural issues, changes, and some confusion in state legislation (Gaceta Parlamentaria, 2002). Within indigenous normative systems, the cargo system represents a complex form of local self-government and reflects the way in which peoples identify the community or collectivity as the center of daily and cultural life. Medina (1996) mentions that the cargo system is a political-religious institution that has characterized a large part of Mesoamerican indigenous life, which is made up of a certain number of positions, or community responsibilities, recognized and respected by the members of the community. The importance of regulatory systems in the Oaxacan context lies in the fact that 73 percent of the municipalities are governed by this political regime, where the sociocultural factor stands out, since the forms of organization and culture, including political and governmental institutions, are observed in a model of society, determine the visions of local development. In the context of the native peoples, through the assembly, collective decision-making is carried out, with the government and the population interacting directly, then it is possible to say that there is a direct democracy, since the citizens and residents are important actors in the solution of collective problems, by expressing directly in the institution called assembly, their needs, complaints, demands, disagreements, and opinions for the solution of problems. The native peoples usually present a high sense of identity towards their municipality, thus weaving a scenario of social cohesion, where the population stays together, to face the problems that arise in the local territory, and to fight externalities. In a modern world, generally Westernized, where individualism, transculturation, and Western hegemony permeate, social cohesion serves as a protective element of the essences of the original peoples. Parsons (1961, in Institute of Legal Research, 2021) points out that, "An individual is socially integrated when he internalizes the norms and values (structures of expectations) that regulate his action in congruence with the social position occupied," which is also related to the sense of identity that the community gives to the individual and the importance that the municipal government gives to each person, since there is a close relationship between the government and the governed. Social cohesion includes aspects such as internalized values, economic equality among community members, participation in cultural events and public decisions, including the political, economic, social, and cultural areas of the municipality. Social cohesion is manifested in personal interrelationships, union mechanisms, such as the community assembly, community committees, social organizations, which generate a sense of belonging to the community. According to the Council of Europe (2005 in Institute of Legal Research, 2021) social cohesion is: the ability of society to ensure the well-being of all its members, including equitable access to available resources, respect for human dignity, diversity, personal and collective autonomy, responsible participation and the minimization of disparities social and economic in order to avoid polarization. It is then about social union, a fabric that seeks the well-being of individuals in the community, in the social, economic, political, cultural, and environmental spheres, with a view to achieving socioeconomic equality of the members of a community. This community attribute works as a protection element against external agents that tend to destabilize the local environment, as in this case the Covid-19 pandemic. With the contribution of Tironi (2010 in Instituto de Investigaciones Jur<unk>dicas, 2021) social cohesion implies a sense of belonging of the members towards their community, they accept and abide by the rules and values developed in community. Cohesion thus encompasses the capacity of a society to generate conditions of union, respect, and collective well-being, seeking equal opportunities to access services and guaranteeing respect for social, economic, political, cultural, and environmental rights. In the original municipalities, there is usually a significant degree of social cohesion, manifesting itself in the sense of identity and belonging of the members of the community, in addition, equality between people is sought, allowing there to be no marked differences between social sectors. In an ideal context, this characteristic gives rise to the members of the municipal government, before exercising the mandate in an authoritarian manner, to preserve the sense of belonging to the social group that elected them, seeking the general welfare. In this sense, social cohesion was one of the elements that The leadership of the Oaxacan municipalities against COVID-19: The case of Santa Catarina Cuixtla, México, an analysis from the crisis management approach allowed the development of effective strategies to contain and confront the Covid-19 pandemic in some Oaxacan municipalities with native populations. According to information from the National System for Epidemiological Surveillance of Respiratory Diseases (SISVER), by January 2021, the second year of the pandemic, 102 indigenous municipalities in the country had remained free of infections, 80 per cent in the Oaxacan territory. "Only four entities -Oaxaca, Chiapas, Puebla, and Veracruz-present indigenous municipalities that do not appear as the municipality of residence of any covid-19 case registered in the SISVER" (Mat<unk>as, 2021). As part of the sense of belonging, neighborhood, and community, from the indigenous worldview there is a sense of social responsibility, which translates into collective care, above individual well-being, a characteristic attribute of native communities. As an example of this, in Villa de Talea de Castro, a Zapotec municipality in the Sierra Norte de Oaxaca, self-isolation was established by shared decision between the municipal government and the population, as a barrier to prevent the arrival of the coronavirus: "In this Instead, the community assembly and the municipal authority determined by a majority vote of their representatives the lockdown of the residents" (Rodr<unk>guez, 2021), thus manifesting the general concern for the collective well-being in the face of the health contingency, in such a way that the cohesion and the sense of vulnerability led the population and the municipal authority to declare self-isolation. In this context, the orographic conditions and the scarce sanitary infrastructure that permeates the entity are added. In this regard, Montes (in Mat<unk>as, 2021) states that: In the native peoples there is awareness and respect for these community decisions, even if they are harsh, there is a great maturity to safeguard the population, they have even suspended their community festivities (because) they know that since there is no hospital network and the proper medical structure in the region the panorama could become tragic. With confinement measures such as the mandatory use of face masks, a sanitary filter, among others, the population of Talea de Castro managed to avoid contagion within the territory until March 2021, more than a year after the start of the contingency: "No one who is not a neighbor or resident in the locality [...] can enter the town. Only the locals could go inside the municipality crossing the sanitary filter, where their medical behavior is recorded" (Montes in Mat<unk>as, 2021). The native populations developed various strategies and actions to avoid contagion, in some cases limiting the population's departures abroad, with the aim of safeguarding public health. This municipality was included in a list of 22 localities in Mexico where, according to the National Council for Political Evaluation of Social Development (CONEVAL, 2021), more than a year after the start of the pandemic, no deaths from Covid-19 had been registered, nor cases, situation that allow to observe the effectiveness of the decisions made by the people. The Indigenous communities are aware of their vulnerability and marginalization from health systems, which is why they have faced the pandemic in their own ways, which in various cases have been effective, both by staying free of contagion and by keeping a low number of infections and deaths from the virus. Cortés (in Ortega, 2021) affirms that, "rural areas are generally more vulnerable with respect to access to health, this was precisely a factor that prevented COVID-19 from arriving", thereby reaffirming the situation of backwardness to which indigenous communities face, which worsened during the pandemic, "these regions and the Indigenous have historically found themselves on the margins of public health, so facing the pandemic represented a particular challenge" The health contingency represents an unknown situation for all the peoples of the earth; however, due to restricted access to health systems, for the original communities it represents a significant crisis situation and a challenge for their local governments. --- Theoretical aspects: The crisis management approach Crises mean challenges, problems, threats, and obligations for individuals and communities, specifically the mission of governments at all levels is to lead society towards optimal living standards, in a context of development and well-being, a scenario in which crises represent obstacles and challenges to achieve such desired scenarios. Christensen, Laegreid, and Rykkja (2013) mention that in recent years more studies have been carried out that analyze crises and their management from the governmental sphere, where the contingencies faced are caused by phenomena such as terrorism or natural disasters (Jann, Jantz, Kühne, & SchulzeGabrechten, 2019). Pandemics or health contingencies also generate crises; thus, the emergence of the Covid-19 pandemic represents a critical situation for civil society, companies, governments, and individuals. As background to this writing Culebro (2021) analyzes: "The nature and characteristics of the organizational/institutional dimensions that influence the coordination and regulation mechanisms during the management of the COVID 19 crisis in Mexico, based on the response of governments local" this study represents an important reference for the analysis of effective crisis management by governments in the country. In a critical context, governments have the responsibility to protect, guide, and organize the population so that catastrophic shocks are as few as possible. In this sense, Boin, Stern, and Sundelius (2016) mention that critical events put people at risk. Social welfare, then crisis management and effective leadership become important elements to give direction and designate responsibilities from the governmental spheres, a premise that coincides with what was expressed by Cruz, Culebro, and Méndez (2021, p. 4) "during a crisis, citizens look to their leaders mainly to try to avoid or minimize its possible damage, and because they serve as a guide to get out of the catastrophe, explain it, and guarantee that such a situation will not happen again". In this context, municipal governments have the duty to act as leaders in their territories (Constitutional article 115), to implement emerging and effective strategies that allow them to deal in the most appropriate way with crisis situations caused by natural phenomena, social movements or health contingencies, as in this case the Covid-19 pandemic. A crisis represents an unknown situation that causes high levels of uncertainty and instability in people, systems, populations, countries, and organizations, as Cruz, Méndez, and Culebro (2020, p. 3) consider it when stating that: "The crisis can be understood as a situation, shared perception, or socio-linguistic construct about a serious threat to the basic structures or to the fundamental norms and values of a system" idea that coincides with ISO22301 where it is stated that the crisis is: "A situation with a high level of uncertainty that affects the basic activities and/or the credibility of the organization and requires urgent measures" (Deloitte, 2021). Cruz et al. (2021) mention that a crisis represents a situation or social construct, which encompasses a threat to the structure, norm, and values of a system, altering peace and order, to which the governments of the three orders must respond immediately and effective to prevent its advance and damage to the population, since "a crisis requires urgent action by the authorities to prevent its materialization, since it generally alters the peace and order of societies" (Backman and Rhinard, in Cruz et al. 2020). In these terms, the Covid-19 pandemic represents a crisis that emerged abruptly at the end of 2019, becoming more acute during the two subsequent years. Boin et al. (2005in Cruz et al., 2021) consider that in a critical situation, governments in general and public servants in particular play a fundamental role in managing the crisis, since the population seeks answers, help, guidance and measures that allow it to be safeguarded, so as not to be significantly affected, in this context, municipal governments should take on the role of leaders to face and overcome crises, in this case the pandemic: "citizens, in order to avoid, or at least minimize, its threat and possible damage, they look for their leaders, be they politicians, public administrators [...] because they must serve as a guide to get out of the crisis..." (Boin, et al, 2005in Cruz et. al. 2020, p. 2). Theoretically Culebro (2021) considers that the Covid-19 health crisis can be analyzed from the crisis management approach, focusing on government action. Crisis management is understood as the main role of a government aimed at strengthening and coordinating the resistance of the population and critical infrastructure networks (Baubion, 2012) and is made up of all the activities aimed at minimizing the impact of the phenomenon, in this case the Covid-19, in the population, the infrastructure and in public institutions. At the local level, those who must exercise leadership to manage the health crisis are the city councils, as the governing body in the municipal territory. The outbreak of the new coronavirus represents a critical context due mainly to three elements present: a scenario of uncertainty in the face of a new disease, since there were no medical care protocols, nor the necessary medications to treat it, which demanded the most effective as possible in the face of the other two elements, the urgency, in this case to stop the spread of the virus in the population and the threat as it is a disease that, depending on various health conditions, can lead to the death of the infected person. From the perspective of Boin et al. (2013), contextualized in Mexico by Culebro (2021b) and Cruz et al. (2020) the effective management of a crisis can be evaluated through leadership in the implementation of ten executive tasks: early detection or early recognition; creation of meaning or giving meaning to the crisis; critical or strategic decision making; horizontal and vertical coordination of instances; coupling or decoupling to the crisis; construction or generation of meaning of the crisis; communication of measures, strategies and results; accountability for performance; learning lessons and improvement of the resilience of the government and the municipality in the face of the crisis, in this order of ideas Boin et al. (2005 in Cruz, Méndez andCulebro, 2020, p. 3) argue that: This is not an easy task as it requires a large, complex and saturated operation, which involves many organizations from both the public and private sectors, as well as supervising the development of the crisis, maintaining communication with key actors, investigating what went wrong, be accountable for actions, create initiatives to improve and re-establish a sense of normalcy, all this in light of the continuous evaluation of the leaders and leadership of different actors. Effective management of a crisis begins with the early recognition or recognition that a threat has arisen (Boin et al., 2013), in this case those who have to share that recognition first are all the councilors of the city council, that is to say the regime. Afterwards, information about the emergence of a crisis must be shared with the administrative apparatus that collaborates with the municipal government, and in a third moment, communicate to the population about the emerging threat, which requires immediate attention. Often a crisis can be foreseeable (Boin et al., 2013), for this analysis it is considered that the Covid-19 pandemic effects could have been visualized in the Mexican context in the 2020 early, when the first cases in China were announced. There are two factors that stimulate the forecast of a contingency: experience in crisis management and organizational development, in this regard Culebro (2021) argues that early detection or early recognition of the crisis is necessary, based on the information provided by the environment and the first signs of an unusual event, in this case the municipal governments had to turn on the alarm signals, based on the communications and guidelines established by international, federal and state organizations, at the end of 2019 and the beginning of 2020. Early detection allows the creation of possible scenarios and the design of strategies to deal with the phenomenon in advance. The second executive task is to create a sense of the crisis, that is, to understand, assume, and internalize its importance and the possible effects on the local population, this task implies a task of raising awareness and political will to recognize the magnitude of the event, based on this possible containment and solution alternatives are generated. Subsequently, the city council must make certain critical decisions (Culebro, 2021) prioritizing the management of the pandemic in the local territory, in the face of other existing public problems. This is a complicated stage (Boin et al., 2013) since it is an unknown phenomenon, in view of which the municipal government as leader must promote a collective understanding of the nature, characteristics, consequences, scope, and potential effects of the crisis. Creating a sense of the crisis is important for governments to make informed decisions, with common sense and the appropriate dimension of the problem. Boin et al. (2013) consider important to establish criteria to process information about the crisis, share it with the right people, consider their comments and feedback, create scenarios about possible futures, and consequences, in addition to visualizing information requirements for decision making. So, both the government and the municipal population must build a meaning by dimensioning the effects of the pandemic in the health, social, economic, political, and environmental fields, to face uncertainty and envision effective strategies in the short, medium, and long term, in order to lessen the consequences. The third action is making critical decisions, crisis management is carried out from the strategic level, that is, by the political regime of the municipal government, the councilors, who are responsible for managing the situation, so making decisions is a task important and crucial (Boin et al. 2013). Strategic crisis managers, in this case municipal governments, should only concern themselves with strategic issues, avoiding focusing on operational decisions. Attention

In the state of Oaxaca, Mexico, two regimes of social and political organization coexist, political parties and indigenous regulatory systems (SNI), the latter present in 73 per cent of Oaxacan municipalities, a context where the main characteristic elements are the community assembly, the cargo system and tequio. During the Covid-19 pandemic, some municipal governments of the SNI regime have stood out for their leadership in managing the health crisis, innovation and implementation of actions to contain and minimize effects in their territories, and prioritizing the protection of public health. In this sense, the objective of this communication is to analyze the strategies that the Oaxacan SNI municipal governments have implemented to face the Covid-19 pandemic, through the crisis management approach. The case of Santa Catarina Cuixtla, Oaxaca, whose socio-political organization system preserves its uses and customs, highlighting features of social cohesion, is presented. The management of the pandemic is characterized and analyzed from ten executive tasks for effective crisis management (early recognition, creation of meaning, decision making, coordination, accountability, learning, resilience, among others) identifying innovation actions, participation, and social cohesion.

a collective understanding of the nature, characteristics, consequences, scope, and potential effects of the crisis. Creating a sense of the crisis is important for governments to make informed decisions, with common sense and the appropriate dimension of the problem. Boin et al. (2013) consider important to establish criteria to process information about the crisis, share it with the right people, consider their comments and feedback, create scenarios about possible futures, and consequences, in addition to visualizing information requirements for decision making. So, both the government and the municipal population must build a meaning by dimensioning the effects of the pandemic in the health, social, economic, political, and environmental fields, to face uncertainty and envision effective strategies in the short, medium, and long term, in order to lessen the consequences. The third action is making critical decisions, crisis management is carried out from the strategic level, that is, by the political regime of the municipal government, the councilors, who are responsible for managing the situation, so making decisions is a task important and crucial (Boin et al. 2013). Strategic crisis managers, in this case municipal governments, should only concern themselves with strategic issues, avoiding focusing on operational decisions. Attention should be directed towards effective strategies that can contain and lessen the crisis, in this regard Heifetz (1994 in Boin et al., 2013) mentions that effective crisis leaders organize an adaptation process, they do not focus on seeking technical solutions. In this task, micro-management must be avoided, where the political regime must be in charge of establishing strategic decisions, and the administrative regime is in charge of operationalizing them at a technical and operational level. The next executive task corresponds to vertical and horizontal coordination (Culebro 2021). Vertical coordination is aimed at the municipal government joining its efforts to contain the pandemic with the federal and state governments, seeking the implementation of effective strategies between the three levels of government. At a horizontal level, the city council must show a close collaboration between its members, and then with the administrative system. Coordination also implies joint work with the social and economic sectors that interact in the municipal territory, through social participation mechanisms. Managing a crisis requires intense cooperation between a variety of organizations and actors who have not previously worked together (Boin et al., 2013). Working in isolation is not effective in a crisis, as authoritarian management of the crisis and excessive control tend to be counterproductive. According to Boin et al. (2013) and Culebro (2021) there is an important self-organization capacity of the affected communities, where local organizations and citizens can be significantly adaptable, creating nodes of cooperation to overcome the critical situation. In a pandemic context, leaders, more than controlling the situation, must be oriented towards coordination, decentralization, and delegation. Then crisis managers should be oriented towards coupling and decoupling of the situation. Decoupling consists of the leaders isolating the problem or the nodes most affected by the crisis (Boin et al., 2013) turning off the parts of the local system that require analysis, allowing the other parts to continue working. Coupling consists of adaptation to the conditions that the pandemic context poses, in this case the functional parts of the system continue to carry out their actions, but receive support or power to be able to face the new environmental conditions. The sixth executive task corresponds to generating the meaning of the crisis, according to Culebro (2021) and Boin et al. (2013) in an emergency situation, the population in general expects from their leaders, in this case the municipal governments, an interpretation of the situation, and above all they seek to listen to solutions and strategies aimed at returning the situation to a state of normality. So, the citizens expect their municipal authorities to provide hope and confidence that the situation can be neutralized. If governments fail to generate trust in the population, chaos will be created, allowing speculation and uncertainty about the future, thereby hindering efficient crisis management and delaying solutions. The creation of meaning is important for the construction of legitimacy of public institutions (Boin et al., 2013), the security, tenacity, and efficiency in the actions of local governments, have a direct impact on the consequences of the crisis. The next essential task in management is communication, in a critical context, communication in all senses and levels is important, as long as it is objective information, based on hard data such as statistics, projections and prospects, because according to Boin et al. (2013) a subjective communication or with false data is one of the main sources of problems. Lack of communication or poor communication can lead to serious repercussions for the safety of people. Horizontal communication includes the type of communication both social and political, from the city council to the population, in this sense mass communication is effective, which must explain the crisis, its consequences and actions to minimize the effects. Through communication, the municipal government must also offer practical advice, explaining to the population actions that must be followed, who must do them and why. The context of a crisis is confusing, fast, and turbulent, for this reason there are difficulties in verifying the information, so only real situations and objective information should be communicated to avoid generating panic in the population. Accountability for performance is the eighth task of leadership in a crisis, carried out before the population, informing in a timely manner about the entire pandemic context both in the municipal territory, and in other latitudes, in such a way that it is communicated at first to society and then to other government agencies on progress, crisis management, investments, and results. Crisis managers have the responsibility to explain what was done before and during the crisis, the reasons for such actions, as well as recovery plans (Boin et al., 2013& Culebro, 2021b), so the municipal government must explain both the gains and the losses arising from the situation. Accountability not only satisfies the legal requirements of government action, it also allows the restoration of confidence in the functioning of public institutions. Governments often find this task difficult, as they focus more on avoiding blame than on generating credibility and legitimacy, in such a scenario accountability is affected by the politicization inherent in a crisis, so governments must establish strategies of transparency and accountability. The ninth executive task is learning, the identification of lessons learned from the experiences lived, the decisions made, and the results observed (Culebro, 2021b), in order to build a bank of information that can be useful in managing future crises. Each critical situation is unique and different, so leaders are required to have the ability to improvise, discover, and experiment, but the ability to learn from the crisis is also needed (Boin et al., 2013). This ability serves to distinguish what is working during the crisis and what worked after it. Learning is required to adapt to subsequent contexts, through the lessons generated, dysfunctional processes are corrected, and solutions are provided. Learning is a complicated process, it basically follows the logic of trial and error, where different actors and levels are involved, such as the human factor, organizational culture, regulation, and system characteristics. Effective learning takes time and requires a culture of research on data and behavior of the crisis, so city councils should allocate an area in charge of collecting information, processing it, and storing it, to be a source of information for future critical situations. The tenth executive task in crisis management is resilience building. According to Boin et al. (2013) a prepared, experienced organization can better manage and contain a crisis, so from the municipal institutional development approach, municipalities must have a certain degree of flexibility and ability to adapt to negative situations. Effective crisis management is carried out by resilient and strong organizations, which can absorb shocks and recover quickly. For building resilient municipal governments, it is important that they engage in preparatory practices such as vulnerability analysis, drills, scenarios, exploration, and networking exercises, which foster a culture of awareness and a shared belief that everyone involved is ready to face some crisis, it means the resilience of a city council as an institution, overcoming the crisis in the best way, and that the variables of the environment do not affect its organizational structure, a scenario where institutional development plays an important role in maintaining the integrity of the government. This resilience is transferred to the entire municipal territory, so that the population can recover from the effects in the shortest possible time. These executive tasks are distributed in three stages: preparation, response, and learning. Early detection, making sense and constructing meaning correspond to preparation; making sense, critical decisions, coordination, constructing meaning, communication, coordination, and accountability are located in the response stage; critical decisions, coordination, accountability, learning lessons, and resilience building are found in the learning phase. This set of strategic actions does not represent a panacea to mitigate critical situations; however, from various studies (Boin at al., 2013;Culebro, 2021 &Cruz, Méndez andCulebro, 2020) it has been shown that following these ten executive tasks it has enabled governments to deal effectively with crises. --- Research findings and discussion: Crisis management by Covid-19 in the municipality of Santa Catarina Cuixtla The following lines analyze the leadership actions from the crisis approach, carried out by the municipal government of Santa Catarina Cuixtla, located in the Sierra Sur region of the state of Oaxaca. This municipality is governed by indigenous normative systems, a regime in which the uses and customs of the population are preserved and manifested in the social and political aspects of collective life. In the management of the pandemic, significant social cohesion and governmentpopulation coordination were observed to deal with the health contingency due to Covid-19. In the early detection task, the municipal government learned about the declaration of the pandemic by the media (news on television and notes on social networks), formally through a statement from the Secretariat of Health of the State Government, during the first days of March 2020. Subsequently, the city council received a notification from the federal government, from which a town hall session was held to analyze the situation and discuss the strategies to be implemented to avoid contagion, in this regard the mayor expressed that in the early stage of the crisis they internalized that they had to protect the population of the municipality "we had to safeguard our people, our citizens, the people of the community" (G. C. Jiménez, personal communication, October 17, 2021). At the local level, the city council took action based on the information received from abroad, once they were formally notified of the start of the crisis by the other two orders of government. The first action of the councilors was the interpretation of the situation, deciding as a first measure the establishment of a sanitary filter at the entrance of the population, since the risk of contagion in the population was visualized due to the daily movement of people from the municipality (merchants and workers) to the district capital, the city of Miahuatlán de Porfirio D<unk>az, as the health councilor referred to it, "we talked with the council and put on a filter, we began to sanitize (sic), first of all, as we did not know how it came the situation we began to sanitize (sic) with chlorine diluted with water to the units (vehicles) that came from Miahuatlán" (L. G. Ram<unk>rez, personal communication, October 19, 2021). The sanitary filter was established on Friday, April 03, 2020, a date on which there were still no cases of Covid-19 in the district of Miahuatlán, so it can be deduced that there was an early detection of the crisis at the local level, establishing this first containment measure. The mayor expressed that from the moment they received the first official communications about the declaration of the pandemic, the city council mobilized, "then we began to agree on how to deal with this situation, the filter was raised [...] during the time the filter was in place, there was no contagion in the community" (G. C. Jiménez, personal communication, October 17, 2021). Given the magnitude of the pandemic, the city council organized itself early to carry out different actions focused on preventing the spread of the virus in the population, including the health filter, the establishment of rules for traveling to other populations, the constant issuance of recommendations to the population through loudspeakers and social networks, the restriction of entry to the municipality for people with a different residence, including the suspension of vehicular traffic to the outside for two weeks, establishing community self-isolation. Another of the measures established to prevent the entry and exit of people to the municipal territory was the establishment of a local community flea market, so that people stocked up on the necessary food, according to the information provided by the councilor of finance and education. "It was an agreement [...] the councilors took care of it, every Sunday, to prevent them from going out on Monday" (B. L. Ruiz, personal communication, October 20, 2021). In México, the tianguis is a traditional market, which takes place in a specific day of the week, in Miahuatlán de Porfirio D<unk>az it takes places every Monday. Regarding the creation of meaning of the crisis, the pandemic represented a challenge for the city council and an obstacle at the same time, since the beginning of the pandemic was conceived during its first year of government, the period includes January 1, 2020, to December 31, 2022. The mayor stated that the projects and works scheduled for the first year could not be carried out within the established time, since various procedures were stopped in the other orders of government, derived from the confinement. In economic terms, the municipal government also identifies that the pandemic has generated important repercussions on the population, reflected in monetary losses, since the majority of the inhabitants depend on the commercial activities carried out in Miahuatlán de Porfirio D<unk>az and on the monetary transfers they receive from the relatives who work in other cities of the country such as Monterrey, Nuevo Léon, Mexico City, Huatulco, Puerto Escondido, Oaxaca de Juárez, among others. The economic depression caused by the crisis generated some tensions and differences between the municipal government, merchants, and carriers, mainly due to the confinement measures established by the city council (confinement, shutdown of shops, transportation service shutdown, closure of the entrance to the village, prohibition of traditional festivals and house parties), affecting the economic income of merchants, public service providers, and employees. The generation of sense of the crisis from the city council was significant, in this regard the councilor of finance and education expressed the following, "we gave a lot of importance (to the pandemic) especially to the care of citizens, protecting everyone's health, we would have wanted that no losses (deaths) have been recorded" (B. L. Ruiz, personal communication, October 20, 2021). It is then identified that there was a high sense of importance towards the health crisis, focusing the efforts of the city council towards the protection of the population's health; however, the crisis also meant a challenge for the municipal authority, since it was an event atypical: "the pandemic was a challenge for us, we were also afraid at the time, because we did not know everything, we had no information about it" (B. L. Ruiz, personal communication, October 20, 2021). The health and finance councilors stated that from a meeting called by the state government they identified the magnitude of the phenomenon "in the first meeting [...] we saw the seriousness (of the pandemic) it was a challenge to face this situation, but we decided to act, from the filter, take care of people, support people" (B. L. Ruiz, personal communication, October 20, 2021). The creation of a sense of the crisis in the municipality of Santa Catarina Cuixtla was carried out through the internalization of information visualizing the magnitude of the pandemic. The councilors held constant meetings and town hall sessions to make decisions aimed at containing the crisis. The meaning given to the contingency was high, manifesting itself in the established actions, promoting confinement in the population and the self-isolation of the municipality, allowing the entry and exit only of the premises that had to move to other places due to urgent situations. Strategic decision-making was exercised from the political regime, that is, the members of the city council, who outlined the main guidelines to deal with the phenomenon. The mayor convened council sessions to analyze the situation and jointly make critical decisions, that is, the sessions represented the spaces for strategic decision-making. As it is a small municipality with 1,495 inhabitants (DATAMEXICO, 2021), the administration staff is small, only made up of the municipal secretary and her assistant, the municipal treasurer, and the police force, so the operationalization of the decisions was executed mostly also by the city council, with the help of the municipal police, made up of 24 elements in total, but every week only six elements are active. Specifically for the operation of the sanitary filter, a councilor and two elements of the municipal police were in charge, in shifts of eight hours. The mayor considers that in financial terms, the cost of facing the pandemic has not been high, since some expenses were made for the acquisition of accessories and necessary equipment "it has been invested, some devices were bought to sanitize and sanitizer (liquid) to support citizens, face masks and gel were bought, the health center and schools have been supported" (G. C. Jiménez, personal communication, October 23, 2021). In the same way, the city council has supported with the purchase of oxygen tanks, the availability of the ambulance and provided pantries to people who have suffered from the disease. The pandemic led to a refocusing of efforts, since the works, social actions and investments had to be stopped in the first months, so that the city council could fully focus on dealing with the pandemic. For coordination in horizontal terms, the city council carried out tasks first internally, delegating responsibilities to all the councilors (political regime), the secretaries, and municipal police force (administrative regime). The organization with the population was carried out through loud communication and social networks, to avoid going to the nearest population center. From the municipal government, the constant issuance of recommendations to the population has been maintained, so that the necessary sanitary measures are followed, and the holding of meetings and parties was avoided. The councilor of the treasury stated that "the colleagues (councilors) went through the streets with the megaphone recommending care, avoiding going out, taking care of the elderly" (B. L. Ruiz, personal communication, October 24, 2021). In the same way, the city council coordinated the population to carry out sanitizations in homes, walking the streets, this action was carried out three times during the months of May, July and September. In the words of the mayor, the majority of the population corresponded to the coordination and instructions issued by the city council, "ninety percent agreed with the measures, and ten percent did not agree, the majority of people continue supporting the decisions of the authority" (G. C. Jiménez, personal communication, October 25, 2021), then there has been a legitimization of the population towards the actions undertaken by the municipal authority. Within the framework of horizontal coordination, during May 2020, a self-organization manifestation of citizens emerged, manifesting some features of social cohesion, when an event occurred that mobilized the majority of the inhabitants of Santa Catarina Cuixtla, a family originally from the municipality but with residence in Mexico City wanted to enter the population to carry out the burial of a corpse, in the face of such an action, the population prevented entry, since the people came from an important source of contagion by Covid-19 in the country. This event made national news, Andrew (2020) wrote about it "Inhabitants of Santa Catarina Cuixtla, Oaxaca, prevent the burial of a woman who died in CDMX". Citizens organized themselves to request the mayor to prevent access to people, since there was a fear that they would carry the virus and it could spread in the population: The leadership of the Oaxacan municipalities against COVID-19: The case of Santa Catarina Cuixtla, México, an analysis from the crisis management approach The municipal authority explained that based on its legal powers, decreed sanitary measures and with the support of the citizens, an agreement was made not to allow access to this community of people from our municipality from the Valley of Mexico to bury a relative (Juárez, 2020). Feeling vulnerable, the residents of the municipality came together to protect themselves from contagion, a situation that evidenced the social union and a notorious capacity for self-coordination among the residents, in this case to safeguard their health "In an assembly in the middle of the road, the mayor, backed by residents, prevented the passage of the body" (Andrew, 2020). With these facts, an important horizontal coordination between the municipal government and the inhabitants is visualized to protect the health of the population "I want to thank the entire population of the Municipality of Cuixtla since they have complied with the regulations of the filters that have been put in place" (Andrew, 2020). As a result of what happened, the inhabitants organized themselves to support the municipal authority to reinforce surveillance in the sanitary filter, integrating groups on social networks (mainly by text messaging) to communicate with each other. With regard to coordination with the economic sector, the city council issued circulars so that health recommendations are followed in the town's businesses, also reducing customer service hours; according to the councilors, in most of the businesses in the town the instructions issued by the town hall were followed. However, in this context of coordination, the city council stated that some people in the population did not give greater importance to the pandemic, expressing an exaggeration in the measures issued by the municipal government, acquiring a greater meaning when the first infected in the territory. In the vertical sense there was no effective coordination between the municipal government and the other levels of government, federal, and state, the health councilor commented that, "from the Ministry of Health, the General Secretary of Government, we received several notifications (about measures) they were nothing more than pure notifications, here we organized ourselves for protection activities, but the actions were ours" (L. G. Ram<unk>rez, personal communication, October 27, 2021). In this sense, at least in the municipality of analysis there was no significant intervention by the federal and state governments for the prevention and mitigation of the effects of the pandemic, their action was limited to the issuance of indicative notifications. During September 2020, the city council requested the support of the Ministry of Health, to carry out a fourth sanitization with the humidifying equipment provided by said body, but it cannot be considered an effective coordination action. Regarding coupling and uncoupling, it was found that the city council promoted the isolation of Covid-19 cases from the rest of the population, as a measure to prevent the spread of the virus. However, they also admit that some relatives of those infected, having been in contact with an infected person, did not respect the quarantine. During the deaths, the members of the city council carried out the excavations of the graves for the corpses of the people who died of the virus, using the backhoe equipment, in order to avoid further contagion. Through isolation, we sought to prevent the spread of the virus, all areas of the city council worked to face the pandemic, headed by the mayor, the councilor argued "where Tino goes (referring to the mayor whose name is Constantino) we all go" (B. L. Ruiz, personal communication, October 26, 2021) thus observing the close collaboration between the members of the city council as parts of a system. The areas of the city council that received the support and feedback were the municipal presidency and the health council, as the main responsible for facing the pandemic, but the councilors stated all the members of the city council have collaborated proactively. In this sense, the city council states that it has adhered to the measures required by this new normality, reprioritizing the actions to be carried out and attending to the health emergency. The city council states that the majority of the population has adapted to the new measures and forms of collective life that the contingency has generated. An important social event that has undergone modifications in the municipality is the celebration of the catholic stewardships, festivities held annually to commemorate the divinities that are venerated in the town, during 2020 the three stewardships and the celebration of Christmas were totally suspended. For the first months of 2021, the celebrations were modified, reducing the number of participants, the call for massive events and the necessary restrictions. The butlers (festival organizing committees) have complied with the measures and the population has reciprocated. During the Day of the Dead celebration, most people followed the recommendations, including the constant use of face masks and disinfectants. In relation to the generation of meaning of the crisis by Covid-19, the members of the city council commented that they had to quickly assimilate the dimension of the phenomenon, since it was totally unknown, it represented a change in the form of political organization, originating a reprioritization of the actions of the municipal work plan, focusing on the protection of the population as one of the main policies of the municipal government. According to what was expressed, the projected works had to wait longer to be carried out: "this situation is quite a challenge, we could not work more on works, we are aware that we take care of the population, before carrying out works, for us the important thing is the citizenship, if we had not carried out these actions, there would have been (sic) more deaths" (B. L. Ruiz, personal communication, October 27, 2021). In this way, it is observed that the health crisis had an important meaning for the municipal government, promoting prevention actions, and acting to help those affected by the pandemic. Regarding the social effects that the pandemic has generated, the councilor for education mentions that "there has been a lot of desertion, with the pandemic they lost their interest in school, there are many children who are lagging behind, there are children who do not want to continue, they have lost interest" (B. L. Ruiz, personal communication, October 27, 2021). It is observed that several children and adolescents in the three educational levels that are taught in the municipality (preschool, elementary and secondary) had abandoned academic activities, since working online is not very stimulating, even some students do not have the electronic means to follow up to virtual classes. This situation represents a scenario of delay in academic content, generating school dropouts. In economic terms, the effects of the pandemic are also evident in the municipality, as the local economy was depressed, some people lost their jobs, or merchants in the city of Miahuatlán saw their sales decrease, however, at the same time, the pandemic meant a encouragement for other businesses to open in the town, mainly fast food, such as taquerias, restaurants and family dining rooms, allowing the population to have some consumption options, mainly with home delivery, thus beginning to reactivate the municipal economy. Some changes have also been generated in the uses and customs of the population, especially in the celebrations of the patron saint festivities, modifying social and cultural coexistence, since the agglomeration of people is restricted. In the same way, community assemblies have not been held, leaving all responsibility for decision-making in the hands of the city council. In terms of tequio, the pandemic has made it possible to resume this action that was born since pre-Hispanic times, by inviting residents to participate in various collective actions, including cleaning the streets and the municipal cemetery. In this same area of the meaning of the crisis, it is necessary to reinforce the information that the city council provides to the population, particularly about the repercussions that Covid-19 has on public health, to continue avoiding cases, because as a result of the application of the vaccine in March 2021, some people have relaxed the protection measures. The city council has established containment measures, mainly the sanitary filter, the prohibition of the entry of people whose permanent residence is not the municipality, the prohibition of crowds, the exhortations to local businesses, the sanitization of homes and streets, among others. However, the city council did not directly try to reassure the population to avoid panic, since this aspect exceeded its framework of action, since the population received information about the pandemic from the mass media, naturally the interpretation of the data was in accordance with the criteria of each person, the action of reassuring the population in a crisis situation corresponds more effectively to the national and state governments, regulating the information disseminated to the population. In terms of communication, the city council stated that it had effective internal communication, given the close relationship between them and their daily coexistence. The main means of communication between the councilors are the town hall sessions, spaces for analysis and discussion of collective problems, they mentioned that the extraordinary sessions to deal with issues related to the pandemic are held constantly. Other means of communication between the councilors are text messages on cell phones and radios, thus allowing information to flow quickly. With regard to external communication, towards the population, it is the mayor and the trustee who coordinate the review of the information that is communicated to the residents; however, all the members of the city council participate in such review, both the owners and the substitutes. The information provided to the population consists mainly of the sanitary measures that must be followed to avoid contagion, the indications on the entry and exit of people to the municipality, the restrictions for the concentration of people and the measures for the celebration of events social and religious. During the first months of the pandemic, the city council transmitted information related to the nature of the virus and the possible sources of contagion. The main means of communication that the municipal government uses to communicate with the population is the loudspeaker in the city hall and in official vehicles, making tours of the population. In addition, social networks, mainly Facebook, have been used to issue official statements. In the same way, the city council mentions that the telephone lines of the presidency and the municipal police are available so that the population can report an emergency or ask for help. The councilors commented that internal communication has been effective, since the key to working in a coordinated manner has been the unity and sense of belonging of the members of the council "Tuesdays, Thursdays and Sundays we all meet, that has helped us, to team up, that has taken us forward, we have not had conflicts between us here, of course sometimes we argue, but things are said here and leaving we are friends as always" (G. C. Jiménez, personal communication, October 28, 2021). Being a small town, there are close family and neighborhood relationships between the members of the municipal government, a factor that has been key to organizational cohesion, which has been reflected in effective results for the population, in terms of pandemic management: "We have remained united, we have not left the health councilor alone, all of this has helped us, we are united" (G. C. Jiménez, personal communication, October 17, 2021). The treasury councilor expressed "teamwork, unity, we are here most of our time, when it comes to work, we are all here, not because we are women we stay in the office" (B. L. Ruiz, personal communication, October 17, 2021) revealing the union and sense of belonging of the councilors. Regarding external communication, the city council commented that communication with the population needs to be further strengthened, although social communication actions have been carried out, it is important to continue informing the community and invite them to continue with the protection measures. From the first moment of confinement, the city council informed and communicated with the population, through printed and digital media, so it is possible to say that the communication has been timely. In the executive accountability task, it was observed that the city council has only reported the actions it has taken to contain the pandemic, however, due to the resulting restrictions, it has not been directly accountable to the population in a community assembly as is customary; semi-annually. During January 2021, an assembly was held, which was very brief, to avoid contagion, the mayor commented that in said event "we invited the people, that whoever had any questions should go to the office or with the treasurer to review the (expense) book" (G. C. Jiménez, personal communication, October 17, 2021). This action reflects the commitment to accountability and openness of public information to the population of the municipality. The treasury councilor argued that social networks have been used to inform the actions carried out and issue security measures, both the federal and state governments have not requested any specific information on the management of the pandemic, in the sense of control. In terms of accountability, the city council has made expenditures derived from the needs that the crisis has caused, such as the purchase of sanitization material and equipment, oxygen tanks, medicines, in-kind support, among others, these expenses are reported to the state congress through the financial reports and accounting statements issued by the municipal treasury. The city council explained that they have suffered a budget cut, specifically from Branch 33 Fund IV Contribution Fund for the Strengthening of Municipalities and Territorial Demarcations of the Federal District (FORTAMUN) "there was a cut in the budget, we received one million one hundred thousand pesos, now we get nine hundred thousand pesos" (G. C. Jiménez, personal communication, October 29, 2021), this situation has affected the adjustment of expenses in items such as infrastructure maintenance and security. At the end of 2020, the city council did not carry out the government report as determined by law, due to contingency restrictions, this action was corrected a month later in the aforementioned community assembly. Regarding the learning that the crisis has left, the councilors mentioned that "working as a team and in unity can do everything, we organize ourselves, due to our uses and customs, many people have contributed their grain of sand, they came to help us, also from part of the citizens who are outside there was a lot of support (B. L. Ruiz, personal communication, October 30, 2021). With these arguments, it follows that the city council was effectively organized to face the pandemic, where teamwork is a key factor for the results obtained. Most of the population has expressed support and shown cooperation with its municipal authority, including people who live in the state of Nuevo León, Mexico City, State of Mexico, and the Oaxacan coast, among other places, have made contributions to collaborate with the expense derived from the pandemic. This collaboration and participation show a significant social cohesion, identity, and sense of belonging of the population and legitimacy of the municipal government. In the area of learning lessons, the city council commented that the implementation of the filter has been one of the most effective measures that were followed to avoid the damage caused by the pandemic, because during the first eight months of the pandemic, when the sanitary filter was working, there was no contagion, this action, as stated, would be implemented again if a similar situation arose. Regarding the actions that would not be implemented again, the health councilor stated, "none of the actions have been in vain, everything has turned out (in a positive way) now we are adapting to living with this disease" (L. G. Ram<unk>rez, personal communication, October 29, 2021) the municipal government commented that the actions they implemented have been useful for managing the pandemic at the local level, since the emergency they had to innovate actions, asking for the The leadership of the Oaxacan municipalities against COVID-19: The case of Santa Catarina Cuixtla, México, an analysis from the crisis management approach collaboration of the population. Regarding the recommendations that it would make to the population for future crisis situations, they mentioned "the population should not ignore the recommendations of the Ministry of Health and the authorities"(L. G. Ram<unk>rez, personal communication, October 29, 2021) because a sector of the population (10 per cent) initially was incredulous about the pandemic existence. The recommendations to the future municipal authorities to face a health crisis revolve around the establishment of a health filter "we could tell the new administration about our

In the state of Oaxaca, Mexico, two regimes of social and political organization coexist, political parties and indigenous regulatory systems (SNI), the latter present in 73 per cent of Oaxacan municipalities, a context where the main characteristic elements are the community assembly, the cargo system and tequio. During the Covid-19 pandemic, some municipal governments of the SNI regime have stood out for their leadership in managing the health crisis, innovation and implementation of actions to contain and minimize effects in their territories, and prioritizing the protection of public health. In this sense, the objective of this communication is to analyze the strategies that the Oaxacan SNI municipal governments have implemented to face the Covid-19 pandemic, through the crisis management approach. The case of Santa Catarina Cuixtla, Oaxaca, whose socio-political organization system preserves its uses and customs, highlighting features of social cohesion, is presented. The management of the pandemic is characterized and analyzed from ten executive tasks for effective crisis management (early recognition, creation of meaning, decision making, coordination, accountability, learning, resilience, among others) identifying innovation actions, participation, and social cohesion.

. Regarding the actions that would not be implemented again, the health councilor stated, "none of the actions have been in vain, everything has turned out (in a positive way) now we are adapting to living with this disease" (L. G. Ram<unk>rez, personal communication, October 29, 2021) the municipal government commented that the actions they implemented have been useful for managing the pandemic at the local level, since the emergency they had to innovate actions, asking for the The leadership of the Oaxacan municipalities against COVID-19: The case of Santa Catarina Cuixtla, México, an analysis from the crisis management approach collaboration of the population. Regarding the recommendations that it would make to the population for future crisis situations, they mentioned "the population should not ignore the recommendations of the Ministry of Health and the authorities"(L. G. Ram<unk>rez, personal communication, October 29, 2021) because a sector of the population (10 per cent) initially was incredulous about the pandemic existence. The recommendations to the future municipal authorities to face a health crisis revolve around the establishment of a health filter "we could tell the new administration about our experience as a suggestion, it is very feasible, the objective is the safety of the population was our focus, caring for citizens" (L. G. Ram<unk>rez, personal communication, October 29, 2021). The councilors argued that the experience with the filter and the sanitization of the population, as well as taking care of the confinement, were effective in containing the entry of the pandemic for several months. In the aspect of improving resilience, in terms of strengthening the municipality as a system, the city council stated that having established the sanitary filter is one of the actions that provided the best results, since contagion was avoided during its operation, however, due to Due to pressure from the population, but living in other places, the municipal authority decided to withdraw it in October 2020. The councilors expressed that in the event of a new health crisis, this action should be implemented again and maintained for facing overcoming the crisis. Regarding the actions that have been carried out to recover from the effects of the pandemic, the city council has authorized the establishment of new businesses, verifying that they follow the corresponding sanitary measures or indicating the delivery of products at home. With regard to public investment, the money that was not allocated to socio-cultural events has been invested in public works, including street maintenance and the improvement of educational infrastructure. Another recovery measure has been the extension of the service hours of collective taxis on the Cuixtla -Miahuatlán route. Regarding sociocultural events such as catholic stewardships, parties and meetings, the city council has authorized the celebration, but monitoring the restriction measures. Social coexistence has a positive impact on the emotional state of the inhabitants, who undergo a process of adaptation to the new normality. For future health or other contingencies, the city council has not contemplated carrying out drills, to prevent further damage to the population. In relation to the action of the population in containing the crisis, the city council argued that approximately 90 per cent of the inhabitants have followed the sanitary measures to avoid the spread of Covid-19, they consider that 10 per cent of the inhabitants have shown some resistance, expressing their disagreement with the protection guidelines, directly to the members of the city council or through social networks. With these data, it is inferred that most of the population corresponded and complied with the established measures, thus highlighting features of social cohesion and political legitimacy. --- CONCLUSION The Oaxacan municipal government has historically been characterized by a framework of institutional fragility, presenting various areas of improvement, municipalities often resort to improvisation, strategic planning is unusual, and management results are not optimal. However, there are also innovative experiences that show that political will, organization, and a spirit of service can achieve important results for social well-being. During the Covid-19 pandemic, some municipalities have stood out, taking the leadership in their territories to face the pandemic in the best possible way, safeguarding the integrity and health of their population. Some important cases are the Villa de Talea de Castro and Ixtlán de Juárez in the Sierra Norte of the state, municipalities that resorted to measures such as self-isolation, the establishment of sanitary filters, the taking of rapid samples to detect Covid-19, the request of records specifying the reasons for entering the municipality, among others. In the Sierra Sur, a particular case was the municipality of Santa Catarina Cuixtla, which distinguished itself from the municipalities in the area for its early action, establishing the sanitary filter and promoting confinement in its population. These actions had an echo at the national level as they were news, mainly because its inhabitants, in coordination with the municipal authority, prevented the access of a family from Mexico City, with the intention of burying a corpse, since it was feared that people could carry the virus and cause infections in the population, which has a significant percentage of older adults. The actions and strategies implemented by the municipal government of Cuixtla were analyzed, using the theoretical approach to crisis management proposed by Boin et al. (2013) and the studies carried out in Mexico by Culebro (2021). The management of the pandemic was framed in ten executive activities that reflect government leadership, in this case the city council under study. In the first executive task, early detection, the city council perceived the pandemic in February 2020, during March the phenomenon was assimilated, establishing the sanitary filter in the first days of April, observing a detection and early action on the crisis. At the same time, the city council took the meaning of the phenomenon, organizing and establishing protection and containment measures. In this activity, the city council should have been more incisive and decisive so that the entire population abided by the rules established since the beginning of the pandemic, establishing sanctions for those who did not respect the confinement. With regard to critical decision-making, by not having enough administrative personnel under its command, the political regime was the one who operationalized the strategies, sharing the time between decision-making and their operationalization. The city council could have hired operational personnel to assist in specific activities, because at one point, both the city council and the municipal police force were overwhelmed in their functions and responsibilities. In terms of coordination, there was no joint work between the three levels of government, since the federal and state levels only limited themselves to issuing communications so that the municipal authorities followed the confinement measures. In this regard, the federal and state governments should have exercised comprehensive leadership, coordinating efforts to better manage the pandemic. With regard to local coordination, the initiative of the population to reinforce the sanitary filter and cooperate with the confinement measures could have been capitalized, favoring the delegation of responsibilities. Greater coordination with doctors who provide services to the population is also suggested. For the coupling and uncoupling, the council resorted to the isolation of infected people, this action should be more closely monitored, establishing sanctions for the relatives of the victims who did not respect the quarantine. In the same way, it is suggested that the city council emphasize the transition to the new normality, resuming activities, but taking care of protection measures. Regarding the generation of meaning, both the city council and the majority of the population gave an important meaning to the pandemic, however it is suggested that for future contingencies the city councils carry out the task of reassuring the population to avoid panic or psychosis, in this sense the municipal government can resort to specialists in the medical and psychological area for the implementation of measures of emotional containment and attention to the population. On the aspect of communication, internally the city council manifested an adequate communication between the councilors, however it is necessary to reinforce the channels of communication towards the outside, mainly with the population, the economic sector, the social sector, and the federal and state governments. In this way the results of crisis management will be more effective. A formal helpline for the population is suggested, or efficient use of social networks such as Facebook and WhatsApp. In the field of accountability, the city council has briefly informed the population about the management of the pandemic and the expenses generated, however, it did not comply in time with the annual government report that by law it must present during the first days of December (2020). Although, there is a significant openness of the municipal government on the control of both the population and the higher bodies, it is suggested to use digital media for the preparation of the government report, in this way greater legitimacy will be achieved. The learning of the crisis for the city council is reflected in lessons on effective strategies, it is suggested in this item to document the actions carried out to generate a bank of experiences, which will be useful for future municipal governments that have to face health contingencies or other kind of phenomena. The most effective measure that the city council highlights is the establishment of the sanitary filter, which was able to contain the entry of the virus for several months. Regarding resilience to overcome the health crisis, the city council has established actions to comply with the new normality, a greater involvement in the economic sector is recommended, promoting entrepreneurship in the population for economic recovery, as well as carrying out social activities that have repercussions in the emotional well-being of the population, including the celebration of cultural events, taking care of sanitary measures. Greater attention to the victims of the pandemic is also required, in order to provide medical and psychological follow-up to overcome the consequences of the disease. Another important measure in this area is the strengthening of activities of identity and belonging of the members of the population, such as tequio and social participation in various public activities. It is convenient that the municipal government of Santa Catarina Cuixtla establish measures such as the integration of a municipal civil protection committee, to prevent disasters, since the municipality is located in an area of high seismological risk. In addition, it is necessary to reinforce the training of the health council and the personnel of the health center so that they carry out preventive actions on health contingencies, reduction of the prevalence of chronic degenerative diseases, The leadership of the Oaxacan municipalities against COVID-19: The case of Santa Catarina Cuixtla, México, an analysis from the crisis management approach among other aspects that have repercussions on public health. In an approximation towards an evaluation of the fulfillment of the ten executive tasks of the municipal government, it is possible to conclude that the municipality of Santa Catarina Cuixtla acceptably carried out six tasks: early detection of the crisis, creation of meaning, coupling and decoupling, the generation of meaning, communication, and learning from the crisis. On a regular basis the local government carried out the other four tasks: critical decision-making, vertical coordination, accountability, resilience, representing areas of improvement that can be strengthened to achieve better results. Some of them have been influenced by factors external to the municipal government, such as vertical coordination and communication. Based on this, it is possible to say that, given the nature of the municipality, the institutional limitations and the socio-economic context of the Sierra Sur, the management of the Covid-19 crisis by the municipal government of Santa Catarina Cuixtla is acceptable, since almost two years after the start of the pandemic, 11 people have died, that is, 0.73 per cent, a relatively low mortality rate. Managing the Covid-19 crisis represents an important challenge for municipal governments, since they are immediately responsible for providing public services to their population and for ensuring general well-being, including the protection of public health. Through the crisis management approach, strategies are identified that have generated positive and effective results in the health of the population, protecting people from external attacks. In addition, certain areas of opportunity are identified that local governments can address to improve their management and fulfill their mission as a city council. The crisis management approach proposes ten executive tasks that municipal governments can follow to generate acceptable results, fulfilling their obligation as leaders in local territories, also establishing retaining walls to protect the population. It is not an infallible recipe; however, its effectiveness has been demonstrated in various latitudes, which can be replicated in the management of the crisis generated by Covid-19. Social cohesion as a contextual variable is important in achieving efficient results, since there is a co-responsibility between the municipal government and the population, so that actions are designed, strategies are executed and optimal results are achieved for the sake of social welfare, protection and the resilience of the population. Some of the main theoretical limitations of this article is the elementary approach to social cohesion, regulatory systems and the social capital of the indigenous communities in Oaxaca. Future research could deepen these important issues for understanding the singular management of the crises in these particular contexts. A methodological limitation of this study is that a survey aimed at a population sample was not included. To complement the research in the future, it is possible to contemplate the application of research instruments to the people of the municipality to broaden the panorama and know the social perspective on the crisis management in the community context. Some possible research lines for the future, which can be glimpsed from this study, is the correlation between social cohesion and crisis management; characteristics and aptitudes of the mayor in managing the crisis and deepening more in community participation in the crisis management.

In the state of Oaxaca, Mexico, two regimes of social and political organization coexist, political parties and indigenous regulatory systems (SNI), the latter present in 73 per cent of Oaxacan municipalities, a context where the main characteristic elements are the community assembly, the cargo system and tequio. During the Covid-19 pandemic, some municipal governments of the SNI regime have stood out for their leadership in managing the health crisis, innovation and implementation of actions to contain and minimize effects in their territories, and prioritizing the protection of public health. In this sense, the objective of this communication is to analyze the strategies that the Oaxacan SNI municipal governments have implemented to face the Covid-19 pandemic, through the crisis management approach. The case of Santa Catarina Cuixtla, Oaxaca, whose socio-political organization system preserves its uses and customs, highlighting features of social cohesion, is presented. The management of the pandemic is characterized and analyzed from ten executive tasks for effective crisis management (early recognition, creation of meaning, decision making, coordination, accountability, learning, resilience, among others) identifying innovation actions, participation, and social cohesion.

Introduction Modern development trends have shown that education and the creation of human resources are at the top of the priorities of national strategies and policies of social, economic and technological progress. In the case of Serbia, reconstruction and transformation of education is one of the prerequisites for the overall sustainable development (Government of the Republic of Serbia [GORS], 2005). Investing in education and providing adequate human capital that can adapt to changed circumstances has the character of an investment, and education policy is not just a policy of creating human capital, but it is a part of the overall development policy of the society (GORS, 2005). Action plan for the implementation of the education development strategy till 2020 (GORS, 2015) suggests that some of the key actions for the higher education are: a) Development of programs for increasing the coverage in priority areas through the adoption of programs with established mechanism for monitoring market needs, analysis of annual or perennial labor market needs and the number of graduates, as well as through the establishment of a body which will assess development and market needs and priorities; b) Modernization of study programs in accordance with the needs of the labor market and with the level of achieved scientific and technological development; c) Encouraging the development and offering of quality study programs; d) Developing the interest of potential students to study at the Bachelor level of academic studies. In this paper we focus on Information and Communication Technology (ICT) sector as one of the prominent and fast growing sectors in Serbia which is constantly in high demand for qualified workforce. Different national initiatives and strategies are encompassing requalification of the unemployed in order to drive this sector and within the National strategy for gender equality for the current period (GORS, 2014) one of the defined ac-tions is requalification of the unemployed women for different ICT profiles. In order to continuously fulfil the demand for workforce in ICT sector, the need for women empowerment is recognized as one of the prerequisites for sustainable ICT workforce generation. One of the areas of ICT with significant perspective is data science. All industries with transaction-based processes and industries that use connected devices, mobile devices, mobile applications, sensors, cameras, online shops, or similar, which generate high volumes of data on a daily bases can benefit from application of data science in terms of increased profits and ROI. Data science is well-established field in developed countries. In Serbia it has just recently became an area of increasing importance and interest, and still much work needs to be done to help the public and companies to truly recognize data science as the driving force of innovation in all sectors. However, important steps are taken and results are noticeable on the labor market as we are witnessing an increase in demand for data scientists, machine learning and AI (Artificial Intelligence) related roles. We can expect that with time the number of companies in Serbia that apply data science and machine learning to their business will increase even more. To successfully respond to this change in the labor market, higher educational institutions have responsibility to help this workforce development by creating and empowering individuals with data science skills. This requires a certain flexibility from universities to innovate and adapt curriculums. While developing data science in Serbia and during the peak of productivity in the overall ICT sector, it is equally important to acknowledge the lack of women in technology, as that is detrimental to progression, and to take intentional actions to overcome this issue. Therefore, the problem should be observed in its whole: from the lack of women in the ICT education, through unrepresentativeness of women on ICT jobs, to the gap in salaries. The remaining of this paper is structured in three subsections. The first section is devoted to gender equality and its importance for the regional development, with emphasize on gender gap in ICT sector in Serbia and gender gap at Serbian universities and faculties that educate ICT students. The second section describes the position of data science as a field of studies at University of Novi Sad. The last section provides some findings on what has been and could be done to promote women in ICT, and consequently in data science. --- Importance of Gender Equality for the Regional Development Gender-wise diversification of workforce and teams in general, has certain known benefits which point to the essential need of businesses to include women at all levels. Various studies indicate that women are better at assessing their capabilities and have less issues with suggestions coming from team members (Kuhn & Villeval, 2013; Williams Woolley and Malone, 2011). These threads contribute to overall improved collaboration within teams. According to the research (Phillips, 2014) women are better at problem solving, while authors of (Hunt et Regardless of knowing that benefits of including women in teams and at all levels in companies extend even further from previously stated, an alarming situation is still present in the STEM (Science, Technology, Engineering, and Mathematics) where, according to Boston Consulting Group (Taplett et al., 2010) women make up 25% STEM workers globally, and only 9% are at leading positions in those fields. Representation of women in STEM is characterized as one of the greatest challenges in women's education. The subject of the research is the evaluation of gender equality at higher educational institutions in Serbia with particular focus on the higher educational institutions that educate ICT students, on the one hand, and the state of data science as a prominent field of ICT which is shaping and changing ICT labor market and will contribute to additional broadening of gender gap in the ICT field. In order to overcome this issue new curriculums in the field of data science must be developed and initiatives to attract women to the field must be practiced. To set the starting point, our first objective is to examine the current situation among ICT students and to evaluate the severance of the present gender gap. The other goal is to examine the situation at the higher education regarding the data science curriculums. For this we use official statistical data provided by Statistical office of the Republic of Serbia. --- Gender Equality in ICT Sector In this paper we are focused on the ICT field, as the part of the STEM, given that several facts will contribute to further widening of the gender gap in the ICT sector that will consequently impact the labour market and that could be addressed through different approach to women throughout the education. 1. Not only that ICT organizations recruit less women than man, but the percentage of women who are leaving IT jobs is higher than percentage of man (Ashcraft, McLain & Eger, 2016; Scott & Kapor Klein, 2017). Therefore, organizations are losing women at a faster rate. 2. In the forthcoming years, automation will significantly influence labour market. It is estimated that within next two decades, it will displace 10% of jobs. International Monetary Fund researchers analysed 30 countries and predicted disproportionately higher loses among women given that particularly clerical jobs, which are more often performed by women, will be under the influence of automation (Brussevich et al., 2018). 3. On the other hand, automation will also lead to increase of certain IT and AI related roles which are highly paid (Brussevich et al., 2018). Given the previously said that women are less represented in the ICT sector this will additionally broaden the gap and contribute to denying women access to highly paid jobs. The issue of gender equity is part of a cultural heritage. It is extremely hard to influence traditional perception on gender roles and universities may not have the power to do that on their own. However, educational institutions, at all levels of education not only universities, have a great responsibility to make necessary efforts to attract women to the field and to provide a steady stream of qualified female students. Without that the landscape of ICT sector is unlikely to change and improve. --- Gender Equality at Universities in Serbia According to data provided by Statistical office of the Republic of Serbia (for 2017/18 school year), Serbia in all regions has more female students enrolled at all levels and types of higher education (bachelor, master, doctoral, vocational studies, public and private universities, faculties, or colleges) than male students. Figure 1 illustrates the distribution of students according to their gender in 2017/2018 school year. There was 57% of female students and 43% of male students. Similar distribution is present for previous years as well. Figure 2 illustrates the distribution of students according to regions in Serbia and student's gender for the past three school years. We can see that Serbia has more female students in all its regions and that this imbalance is the least pronounced in the South-East part of the country. Not all fields of studies are equally interesting to female students. Figure 3 illustrates the distribution of students for the school year 2017/2018 according to the field of studies and student's gender. We can observe that ICT field is the least interesting for female students and that women currently represent 27% of Serbia's ICT students. When they enter the business world this number will decrease due to attrition that is present along the career trajectory and the fact that there are large gender gaps from graduation, the entry-level through to the top-executive positions (Taplett et al., 2018). --- Figure 3: Distribution of students in 2017-2018 according to the field of studies Source: Authors based on data provided by Statistical office of the Republic of Serbia Gender gap in ICT education varies through different regions of Serbia. Figures 4a and4b illustrate gender gap at universities in Serbia according to the region. We can see that gender gap at public universities (Figure 4a) is more pronounced in Southern parts of the country. Public universities from Belgrade and Vojvodina educate half of the overall Serbia's ICT students, while the rest of the students is distributed over other educational institutions. It is promising that Belgrade universities have reached almost equal distribution of female and male students. The gender gap is more pro-nounced at private universities, and particularly at professional studies (Figure 5a and5b) than at universities and faculties. Considering that professional higher schools or colleges educate approximately 21% of overall ICT students these numbers are not to be neglected. Based on official statistics, the gender gap is evident at Serbia's higher education institutions. Although universities do not have the sole power to break society's or perceptual barriers regarding ICT sector as a maledominated filed, they should take mitigating steps through support system and role-model representation. This will be discussed in more detail in the following sections. --- The Representation of Data Science as a Field of Studies in Serbia Within ICT, data science as a field of studies is still in a developing phase in Serbia. Some of the first steps that universities can do, is to work on developing data science curriculums to provide steady stream of data science workforce. This position allows universities in Serbia to make gender diversity in data science a priority from the start. Most of the higher education institutions in Serbia, which educate ICT students, offer data science related courses, while study programs fully devoted to data science are predominantly at master level. This leaves tremendous space for further development of data science curriculums. At the University of Novi Sad four faculties educate ICT students. - --- Development of Data Science and Gender Equality at Serbia's Universities Apart from a really important job of universities, to develop necessary curriculums, universities should address two key challenges: 1. To increase the number of female students who are studying data science or ICT in general 2. To develop business climate that will help retention and support of women. The big question is how to achieve these goals. Some of the institutionalized initiatives are: a) requalification of women for ICT within National strategy and action plans for gender equality and a certain number of faculties are involved in this initiative; b) opening IT programs in high schools. Other initiatives are either sporadic or non-institutionalized, while initiatives concerning gender equity most often come from associations, nongovernment entities or projects, or companies devoted to promotion of women in ICT. Potential directions that will allow academic community and higher education institutions to be more involved include, but are not limited to the following: --- Conclusion To assure development of higher education curriculums that are in line with market trends and changed demand in ICT sector, data science should be one of the fields in focus. Although data science is still in a developing phase in Serbia, an expansion of its use in various sectors is expected within the next several years. To proactively address these changes, which are already happening in more developed countries, new data science curriculums should be developed now. This starting position allows universities to set gender diversity a priority from the start. It is essential to promote women in data science and ICT, not only because of financial benefits which are linked to gender diversification of teams, but because of the evident lack of qualified workforce and increased demand for ICT and data science professionals. Gender gap is evident at higher education institution that have ICT study programs and issue an ICT related diploma, and educational institutions should take more actions to promote women in ICT and acquire more women to the field. In this paper we have identified potential starting directions for increased involvement of universities with respect to this important issue.

Information and Communication Technology (ICT) sector in Serbia is fast growing and in constant high demand for qualified workforce. In order to continuously fulfil the demand, the National strategy for gender equality defined as one of the actions, requalification of the unemployed women for different ICT profiles. Investing in education is the policy of creating human capital that influences the overall development of the society. Women empowerment through education is recognized as one of the prerequisites for sustainable ICT workforce generation. Data science represents particularly important field of ICT which is still in its development phase in Serbia. All industries with transaction-based processes and industries that use connected or mobile devices or similar, which generate high volumes of data on a daily basis can benefit from application of data science in terms of increased profits and ROI. While developing data science in Serbia and during the peak of productivity in the overall ICT sector, it is important to acknowledge the lack of women in technology, why that is detrimental to progres-

Introduction The Papanicolaou (Pap) test was introduced in the 1950/60s and has considerably reduced cervical cancer incidence and mortality by up to 80% in Europe (Anttila et al., 2004). A 60% decrease was reported for the United States after the first 20 years of cervical screening (Marcial, 1977) and a decrease of similar magnitude has been reported in the Canadian provinces of British Columbia (van Oortmarssen and Habbema, 1986) and Manitoba (Choi and Nelson, 1986). Cervical cancer arises mostly in women who are not regularly screened (Albrow et al., 2012;Spence et al., 2007) and its incidence is disproportionately high among Indigenous women in Canada, the United States (Vasilevska et al., 2012;Decker et al., 2015), Australia and New Zealand (Jackson and Ward, 2000;Reath and Usherwood, 1998). Barriers to Canadian Indigenous women attending cervical cancer screening or following-up on abnormal results have been described by us (Maar et al., 2013) and other research (Morgan and Wabie, 2012). These include social, historical and cultural determinants of health, such as the impact of colonialism and social marginalization; structural limitations, such as reduced access to health resources; and reduced income and lower educational opportunities in comparison with the population as a whole. In addition the invasive nature of the Pap test has been cited as a barrier for this group (Maar et al., 2013). Indigenous women often report not feeling respected in the health care system (Spence et al., 2007) and many have experienced culturally inappropriate service provision (Read and Bateson, 2009). The intersection of social, cultural and historical disparities of health along with a lack of understanding about the purpose and benefits of cervical cancer screening has been identified as a key factor as to why Indigenous women more often develop cervical cancer and die from it than non-Indigenous women (Morgan and Wabie, 2012;Read and Bateson, 2009). Many Indigenous women do not know what a Pap test entails (Hislop et al., 1996) and the link between sexually transmitted infections (STIs) such as human papillomavirus (HPV) and cancer leads to stigmatization (Jackson and Ward, 2000), further exacerbating insufficient screening attendance. Additionally, some Indigenous populations view death as a natural part of the life cycle. Illness is described as an imbalance, and removing a part of the body [like the scraping of cells for a Pap] is considered harmful (Strickland et al., 1999). Hence, it has been reported in some studies that fear of death may not be a motivator for prevention services and may constitute a cultural barrier to cervical screening (Strickland et al., 1996). A recent review (Clarke et al., 2013) that evaluated medical interventions with ethnic minority women showed that the most common strategy for minority health is education about a disease with focus on the patient as the target, lacking emphasis on community engagement and investigation of the multiple, intersecting levels of the health care system and how they impact screening uptake (Clarke et al., 2013;Morgan and Wabie, 2012;Reath and Usherwood, 1998). Our study examines the engagement of First Nations women from northwest Ontario, Canada in cervical cancer screening. We hypothesize that increased, regular screening attendance and follow-up of abnormal screening results would reduce the higher cervical cancer burden in this population. We have previously reported screening recruitment strategies with First Nations women and, based on interviews with health care professionals (HCPs), we have reported that education is one of the key factors for increasing their cervical screening attendance (Maar et al., 2013;Maar et al., 2014). To obtain a more representative and comprehensive understanding of the views held by members of our partner communities, we extended our study to include community members without formal health education to ensure that their voices would be heard in the research. Here, we present the data from the individual HCP interviews and the focus groups with community women focusing on potential educational strategies with the future aim to increase awareness for and uptake of cervical screening in Canadian First Nations women. --- Methods --- Communities and Study Design We developed a participatory action research (PAR) project, entitled the Anishinaabek Cervical Cancer Screening Study (ACCSS) in collaboration with ten Robinson Superior First Nations communities in northwest Ontario, Canada to address cervical cancer screening in First Nations women (Wood et al., 2014). The demographic characteristics of the partner communities have been described in detail elsewhere (Wood et al., 2014). Our theoretical framework is based on the ethical space concept developed by First Nations scholar Ermine (2007) to create respectful and meaningful communication and dialogue across differing cultural worldviews (Zehbe et al., 2012). Following PAR principles which "aim to democratize the research" and work in partnership with marginalized groups (van den Hoonard, 2015) First Nations' partner communities were included at all stages of the research process: from initial project planning, to sampling, data collection and analysis, as well as decisions about dissemination strategies (Esterberg, 2002). Respecting First Nations' ethical protocols in Canada formal research agreements were negotiated with the Chief and Band members of each of our partner communities after face to face community meetings in addition to obtaining Tri-Council research ethics approval from the Principal Investigator's host institution. ACCSS consists of a qualitative strand with interviews and focus groups, and a quantitative strand with a cluster, randomized two-armed screening trial (Wood et al., 2014). Here, we focus on the theme 'education' of the project phase involving qualitative research. For healthcare provider interviews, we asked the health directors in each community to select one or more key informants among the professional health care staff. For focus group discussions, the community-based health directors were asked to recruit participants without formal health care education who would be in the target group for cervical cancer screening and education. Small incentives (in the form of money) were paid to all participants to compensate for their time, as is accepted practice in this context. During the focus groups, the research team provided either lunch or supper, mostly catered by a local community member. Two researchers conducted the interviews and focus groups (Maar et al., 2013;2014). For focus groups, an assistant helped with audio recordings and note taking during the sessions. The study was approved by the Lakehead University Research Ethics Board (#038 11-12/ROMEO #1462079 and 12-13/ROMEO #1463139). --- Participants We conducted interviews (lasting from 40 to 90 minutes) with 18 experienced, younger (under 40 years of age) and middle-aged (above 40 years of age) HCPs from August 2011 to February 2012 (Maar et al., 2013;Maar et al., 2014). We conducted nine focus groups (lasting from 80 to 120 minutes and including between 3 to 11 women at each community site) from May to July 2012 and then two additional ones in April and May 2014 to reach saturation with the focus group data. The majority of participants identified as being First Nations. Criteria for inclusion in the focus groups included being female resident of the respective community with no formal health education beyond a secondary school level. The lower age limit of participants was 18. Attendees [altogether 76] were mostly adult women but in some cases, female youth and female elders were also present. --- Data Analysis We used semi-structured interview guides for individual HCP interviews and the community women's focus group discussions. We explored questions related to the providers' perceptions on effective ways to reach First Nations women with screening information, potential or promising strategies and special needs to help motivate them to participate in cervical cancer screening. The community focus group discussions were centered around experience with cancer in general and cervical cancer screening in particular and why women do or do not attend and how to educate women about cervical cancer screening in the community. The interview and focus group recordings were transcribed verbatim, reviewed and coded by multiple team members to identify emerging themes for analysis and triangulation. A process of "open coding" (van den Hoonaard, 2015) was used to identify basic level themes which were then grouped into key themes through team review and discussion of the transcripts. At the request of our partner communities no identifiers other than HCP or Focus group participant are provided in the data discussion to preserve the anonymity of the participants. --- Results The Need for more Awareness of Benefits Related to Prevention and Screening Ideally, to promote health and wellbeing and to prevent chronic diseases like cancer, a steady flow of health education provided by researchers to HCPs and directly or indirectly to community members is required to create awareness for prevention and screening as part of the community wellness program. As has been noted in an Alaskan study with American Native women (Stillwater et al., 1995), there is little knowledge about cervical cancer and therefore women do not go for screening. Interestingly, this was emphasized by most HCPs but not by focus group participants: 'I don't feel that they are aware of the screening or if it's just they've put that block up and they don't want to do it, but I don't think that they're educated enough to realize so, you know, if something more can be done is the awareness of screening' (Interview 14). HCPs can play an active role in clarifying the benefits of cervical screening and they should convey evidence-based information about guidelines and service to community women. For this reason, the flow of knowledge transfer between researchers, HCPs and clients is essential. Indeed, as has also been reported by Strickland et al. (1996), researchers need to ensure that new findings are accessible to HCPs and conveyed to community members as evidenced by the interview below: 'You need to make sure that you train the trainers, like, teach, the, the, the CHRs [community health representatives], and the women in the community that are providing service' (Interview 2). As noted by another interviewee and in previous research (Sheridan et al., 2004), by emphasizing that cervical cancer is almost entirely preventable with regular screening, HCPs can share encouraging information with their clients: 'We know that cervical cancer is 100% preventable, and I know because I read stats and see statistics that Aboriginal women are the number one on the list for dying from this. So I think maybe reinforcing that message, that this is 100% preventable' (Interview 4). While HCPs generally emphasized the need for increased awareness about cervical screening, focus group participants wanted to learn more about the cause of cervical cancer, namely HPV. In particular, researchers need to explain statistics about the frequency of HPV infections in humans [e.g. that it is the most widely spread sexually transmitted agent] and the fact that, despite it being the causative agent for cervical cancer, not all HPV infections lead to cancer: 'You have to explain more about that test because it sounds like if you have HPV you are going to get cancer and most women in the communities will be scared' (Focus Group 6). --- Educational Interventions through Public Events As suggested in previous studies (Read and Bateson, 2009;Reath and Usherwood, 1998;Stillwater, 1999;Subrahmanian et al., 2011) community educational events are a key strategy for health promotion and to increase the health literacy of the public. HCPs and focus group participants both thought that health education has to be on a continuous basis to be successful. The majority of the interviewees recommended events for women, which could also be blended into existing programs: 'The mothers would all come and the grandma's would come and the aunts would come so I would have a big number there and I would go through everything and I, you know and I, the team would be prepared and I would have guest speakers come in as well and there would be a dinner, a luncheon or a feast of some sort and I wouldn't cook the feast and I wouldn't have anybody in the community cook it, I would bring somebody in because I find that's most intriguing, that you know, and I would bring somebody in who had a little bit of charisma like I would bring in probably a Tai Chi instructor who cooks meals' (Interview 7). 'If we plugged into some existing programming, we got a parents and tots program or, you know, we set up a circle that was going to have some, or a meeting that was going to have something but there was a lunch involved and you know, we could plug some existing groups or programs into it' (Interview 10). However, and as noted by other research to involve the'significant other' (Gotay and Wilson, 1998), focus group participants stressed that educational interventions for males and females alike are an important strategy, especially to de-stigmatize HPV infection-to not always put the blame on women being the carriers but to inform that men are also carriers of this virus: 'But you're, it's the truth though because if the men are, if weren't educated an hour ago, who's to say they know anything about this [HPV and that both sexes can be infected], which they probably don't. They know about herpes and all that because it's so publicized' (Focus Group 6). --- How to Achieve Success with Educational Events in Schools As we heard during several interviews and as evidenced by the quote below, the partner communities seem to host regular educational events for high school students as part of their curriculum to ensure the students will attend lessons in sexual health: 'Well, we see, generally we start with uh, with uh, in the school. With the health nurse, um, we go up there, we talk about uh, human sexuality, and um, we do, we invite our high school students, and we come, and they come there, and we talk about STDs [sexually transmitted diseases] and transmission and, um, you know, all, all about the birds and bees, you know, they know we provide condoms and there at the clinic. Uh so you know the kids are pretty open to me, like I guess there is trust there' (Interview 4). However, the lack of information that HPV, a sexually transmitted agent, causes cervical cancer as a rationale for HPV vaccination was noted, mostly by focus group participants, implying a need for adjustment in the health education focus in Canadian schools: 'My girl came home from school saying you know there's an HPV vaccination, I said well what do you know about it, she said nothing really' (Focus Group 5). Nevertheless, some interviewees shared with us successful experiences that they have had with their teenage students. This finding supports Read and Bateson's work on achieving positive outcomes of educational interventions ( 2009), e.g. via culturally sensitive gender dynamics, humour and open-mindedness: 'The regular program for girls meet every second Monday and the boys meet every [word missing], and it's called the Moon Time Girls and it's called the Wilderness Boys, and it's just their group' (Interview 3). 'Yeah I think they got it because they still, when they see me on the streets here, we kind of giggle about it because we used bananas, you know, and they'll ask me are you bringing bananas next week, you know, so it was a fun thing' (Interview 6). 'I do the whole spiel, with, STIs you know, including HPV, you know, genital herpes as well as Chlamydia GC. And you know, so the cervical cancer, so they're aware of the need for it, and it's you know, it's a survey and choice, so I go through all that' (Interview 5). --- Not all Community Members Can Be Reached through Educational Events Some of our interviewees and focus group members thought it challenging to entice people to come to educational events but felt that women might appreciate hearing from a cancer survivor or an experienced advocate: 'Maybe a person that already went through it, like you know, I had cervical cancer and obviously beat it, you know, maybe something like that type of speaker' (Focus Group 7). While educational events may attract individuals who normally socialize in their community, we learned, mostly from interviewees, that there are others that shy away from attending: 'We've done lunch and learns on it [cervical cancer screening] but we don't have like, a lot of people that come out, where they share their stories, or, whatever they learned with other community members. We usually have the same people coming all the time, so it's, we're not hitting all the target risk' (Interview 13). One potential reason why health education may be difficult to deliver to adult community members is due to the legacy of historical trauma and social determinants-related challenges, as communicated by several HCPs: 'It's always the same people. Then we thought, maybe, there's, like problems on reserve with pills and stuff like that so, we're thinking those families just don't wanna be involved in anything' (Interview 13). Another reason for non-attendance may be related to age: 'The older generation they are very set in their ways and um, they're a bit more difficult to um, to get through to with regards to health teaching and prevention' (Interview 16). --- Potential Strategies to Engage Hard-to-Reach Women for Cervical Screening To reach women who do not normally attend educational events in health, the use of traditional [pamphlets, radio, TV, advertisement] and new social media [website, Facebook, Twitter], as reported in other research (Christopher et al., 2005) and outlined by several focus group participants and HCPs, could be a useful option: 'I think um, you have to gear some of your uh, your media stuff to make it for certain age groups. Um, like pamphlets, you, make them look more teen friendly or something like that and very short sentences and do points, like to the point and then uh, maybe for the older generation' (Focus Group 9). 'Everybody's on Facebook' (Interview 1). While the above approach is rather impersonal, another way of involving hard-to-reach women would be on a personal face-to-face level. As described in previous investigations with Nganampa Aboriginal women in South Australia (Reath and Usherwood, 1998), urban American Native women in North Carolina (Dignan et al., 2005) as well as Cherokee women in the Northern Plains of America (Subrahmanian et al., 2011) and as suggested by health disparities intervention researchers (Clarke et al., 2013), most of our interviewees and some focus group members thought that developing personal relationships would make a difference in drawing people to educational sessions. Various suggestions, involving the HCP encouraging the client or the HCP going door-to-door, as well as family members or friends and colleagues [peers] reminding each other on a word-of-mouth basis, were provided: 'I think I'd have to go, someone would have to go door-to-door and we'd have like a list of what age group you're looking at and then we can go and pick certain individuals that we know that would probably participate and then get the group together and speak about it' (Interview 9). 'See like my niece, I went for a Pap test there like a couple of months ago and she, I was like go do a Pap test and she was like what is that, and I told her, you know, it could cause this and this, you know, and she's like, um no I don't feel like it. That's my great niece and she has a baby and I told her got to get regular checkups and she's like do you? I was like yeah. She ended up going for a Pap test though' (Focus Group 8). --- Discussion Continuous health education for men and women is needed to steadily develop and maintain a screening culture framed by the idea of women's health and wellbeing rather than by preventing diseases such as cancer. Notably, disease prevention per se is not traditionally part of Indigenous framing of health for many communities. This can only be achieved 'long-term' through an ongoing dialogue which builds trusting and collaborative relationships between Indigenous peoples, researchers and HCPs using well-grounded ethical principles based on Indigenous values. Researchers must facilitate and promote such interactions as well as provide feedback to the public and stakeholders. Stakeholders that are responsible for screening guidelines must be open-minded and willing to adapt them for the benefit of Indigenous peoples as necessary. It seems challenging to reach under-screened women and convince them of the benefits of cervical screening. Gotay and Wilson (1998) argue that women with strong social ties have higher screening rates than women with troubled or undeveloped social networks. Therefore, strategies involving personal face-to-face dialogues between researchers and HCPs as well as between HCPs and their clients need to be pursued to improve the present educational environment and screening habits (Clarke et al., 2013, Dignan et al., 2005;Reath and Usherwood, 1998;Subrahmanian et al., 2011;Zehbe et al., 2011). Combined with interventions generally directed toward improving access to the health care system, which other research has already called for (Reath and Usherwood, 1998), we infer such strategies will reach more women in their communities. Future ACCSS directions include new educational approaches such as arts integration (Russell and de Leeuw 2012) to attract more women to attend screening. Integrating the arts as a teaching method in educational settings has demonstrated increased learning engagement (Burnafurd et al., 2000;Gelineau, 2004). Art making in the research process produces various works, which can then be purposefully used to mobilize knowledge across the audience spectrum and as teaching tools for students in formal and informal educational settings (Cole and Knowles, 2001;Sameshima and Sinner, 2009). Being mindful of the potential stigma around HPV, we note the importance of being transparent that HPV infection impacts most humans at some point in their lives. In addition, the effects and benefits of the HPV vaccination for both sexes should be discussed. Such clarifying teachings can be more effective when attendees participate in embodied learning, e.g. with audio-visual tools and anatomy models (Sameshima, 2008) and substantiated through firstperson stories of survivors, peer teaching and community engagement.

Recognition of the need to decrease cervical cancer rates in Indigenous populations has been ongoing-yet few successful interventions have been reported. In addition, literature addressing the challenges and barriers associated with designing screening programs aimed to specifically reach Indigenous women is limited. Here, we report findings from a mixed methods cervical cancer research project conducted in partnership with 10 First Nations communities in northwest Ontario, Canada. Individual interviews with community health professionals (the majority of whom identified as First Nations) stressed that awareness of cervical screening benefits is lacking. In contrast, focus group participants (women with no formal health education) emphasized the desire to learn more about the science of human papillomavirus (HPV), and that a positive HPV or abnormal Papanicolaou test need not mean a woman will undoubtedly develop cervical cancer. Both the health professionals and the focus group participants highlighted that sexual health education must start early, in schools, preferably before girls are sexually active and that it has to continue throughout life to create a screening culture with a focus on women's wellbeing. Both interview and focus group participants highlighted that sexual health education must start early, in schools, preferably before girls are sexually active and that it has to continue throughout life to

Spain standing out as one of the most seriously hit in economic terms, especially with regard to unemployment, which is a wellknown risk factor for psychological morbidity. 1 A small number of studies, several of which were based in other Mediterranean countries, have investigated the health effects of the recession. For example, Greece appears to have suffered an increase in suicide rates and a deterioration in self-perceived health status. 2 In Spain, the evidence on this issue to date is limited to one study that highlighted an increase in suicide rates in working-age men 3 and a second study that indicated a substantial increase in the prevalence of most types of mental health disorders among primary care attendees between 2006-2007 and 2010-2011, with unemployment, mortgage payment difficulties and evictions emerging as major risk factors, especially for depression. 4 In the present study, we add to evidence of the recession's effects on mental health outcomes in Spain. Importantly, we present evidence not only on average health effects in the population but also on changes in socioeconomic inequalities in health, an aspect of research on economic fluctuations and health that has largely been neglected to date. 5 More specifically, we analyse how mental health and socioeconomic inequalities in mental health changed in the Spanish population between 2006-2007 and 2011-2012, and we also analyse the role that working status has played in these changes. --- Methods We undertook a before-and-after study with control variables, using data from the 2006-2007 and 2011-2012 rounds of the Spanish National Health Survey, a representative cross-sectional survey of the non-institutionalized Spanish population. Data were collected The dependent variable was mental health, which was measured using the 12-item version of the General Health Questionnaire (GHQ-12), a widely used tool for detecting risk of current, diagnosable anxious and depressive symptoms, and which has been validated in the Spanish population. 7 We used a two-point scoring method, rating a problem as absent (0) or present (1). Participants with a score!3 were classified as having poor mental health. The main independent variable was the survey edition, represented by a dichotomous dummy variable (1 = 2011/2012; 0 = 2006/2007). Other variables were two socioeconomic variables (social class and level of education), working status, main breadwinner and country of birth. Social class was assigned according to the current or previous occupation of the interviewee or, if he or she has never worked, the occupation of the head of the household. 8 Interviewees who were the person who contributed most to their household budget were referred to as the'main breadwinner'. To measure socioeconomic inequalities in mental health, we calculated two different measures for each survey year using either social class or education as a socioeconomic indicator: (i) the relative index of inequality (RII), a measure of relative health inequalities; and (ii) the slope index of inequality (SII), a measure of absolute health inequalities. We fitted two Poisson regression models with robust error variance to test for association [PR and 95% confidence interval (95% CI)] between poor mental health and the survey year (2011-2012 vs. 2006-2007) for each sex and for each of the categories of the other independent variables described above. In Model 1 we estimate crude prevalence ratios, and in Model 2 our estimates were adjusted for age and working status. --- Results Descriptive statistics suggests an increase in the prevalence of poor mental health among men (from 14.7 to 16.9% between 2006/2007 and 2011/2012) but not among women, whose poor mental health prevalance has decreased (from 24.6 to 22.7%). Both absolute and relative inequalities in poor mental health increased among men (RII = 1.02, 95%CI 0.99-1.04; and 1.08, 95%CI 1.05-1.11, for 2006-2007 and 2011-2012, respectively; P-value for change <unk>0.001; using social class as the socioeconomic indicator), but remained stable among women between the two survey rounds (although absolute inequalities were greater among women than among men, according to both measures). The change in prevalence of poor mental health among men is confirmed by the results of regression Model I, which showed a 15% increase in 2011-2012 compared with that in 2006-2007 (PR = 1.15, 95%CI 1.04-1.26). This increase was relatively larger in men in the 35-44 (PR = 1.24, 95%CI 1.04-1.47) and 45-54 years age groups (PR = 1.29, 95%CI 1.07-1.55), those in social class IV (PR = 1.23, 95%CI 1.07-1.40), those with primary or secondary education (PR = 1.29, 95 %CI 1.02-1.62; PR = 1.27, 95%CI 1.07-1.49, respectively), foreigners (PR = 1.33, 95%CI 1.00-1.77) and breadwinners (PR = 1.15, 95%CI 1.03-1.28). These associations were not markedly affected by adjustment for age (data not shown), whereas, importantly, none of these associations were maintained after additional adjustment for working status (Model 2). Among women, the prevalence of poor mental health decreased overall (PR = 0.92, 95%CI 0.87-0.98), with the decrease concentrated among those younger than 24 years, the employed and non-breadwinners. --- Discussion In this study, we observed an increase in the prevalence of poor mental health among men and a slight decrease among women between 2006-2007 and 2011-2012. Among men, this increase is attributed to employment status. Socioeconomic inequalities in the prevalence of mental health increased among men, but remained stable among women. The greater impact of unemployment on the mental health of men and less advantaged socioeconomic groups observed in our study is similar to that reported by a previous meta-analysis (which, however, did not focus specifically on times of recession). 1 Moreover, as mentioned above, suicide rates also increased among working age men, but not among women, during the economic crisis in Spain 3 ; this supports our findings because suicide can be considered a proxy for mental health problems. Recently, a study using a similar approach in the UK also found a trend of increasing inequalities in poor mental health during the period 1991-2010; inequalities worsened among men (and did not vary significantly among women), but contrary to our study, employment status did not play an important role. 9 The differences in the trend of the mental health effects among men compared with women, which is most noticeable among individuals with a secondary level education, have been attributed to the notion that work is strongly related to the social role of men as the main household breadwinner, while women find psychological compensation in their family role as substitute for employment. 10 While a full dissection of the various causal pathways behind our results is beyond the scope of this article, our data (not shown) suggest that the improvement in mental health is concentrated among employed women, who may experience additional recognition and greater self-esteem in a new breadwinner role. With regard to other changes that could modify the effects of socioeconomic factors on mental health, it is worth mentioning that public health care coverage in Spain was still universal until 2012, and that a sensitivity analysis with additional adjustment for social support did not significantly affect our results (data not shown). This is the first study to analyse mental health outcomes at the population level before and during the Spanish crisis. As the 2006-2007 survey was the first to introduce GHQ-12, we cannot assess preceding trends. However, the increase in mental distress is consistent with the aforementioned study on suicide rates, which was not disaggregated by socioeconomic position 3, and the study on general practitioner users, which was not disaggregated by sex. 4 Moreover, evidence that increasing unemployment accounts for deteriorating mental health in men supports the plausibility of the role of the crisis. --- Conflicts of interest: None declared. --- Key points This study shows a deterioration in mental health among men during the economic crisis in Spain, especially among those from low socioeconomic position. Among women, we noted a slight improvement in mental health, driven by younger individuals, employed and non-breadwinners. Socioeconomic inequalities in mental health became more pronounced among men but not among women. Social protection policies need to be reinforced, especially among individuals from a disadvantaged socioeconomic position.

We analyse how mental health and socioeconomic inequalities in the Spanish population aged 16-64 years have changed between 2006-2007 and 2011-2012. We observed an increase in the prevalence of poor mental health among men (prevalence ratio = 1.15, 95% CI 1.04-1.26], especially among those aged 35-54 years, those with primary and secondary education, those from semi-qualified social classes and among breadwinners. None of these associations remained after adjusting for working status. The relative index of inequality by social class increased for men from 1.02 to 1.08 (P = 0.001). We observed a slight decrease in the prevalence of poor mental health among women (prevalence ratio = 0.92, 95% CI 0.87-0.98), without any significant change in health inequality.

Introduction Sleman Regency, located in the Special Region of Yogyakarta Province, has become one of the leading tourism destinations in Indonesia. Its natural beauty, cultural wealth, and historical heritage make Sleman Regency an attraction for local and international tourists. However, the rapid growth of tourism also brings challenges and consequences that must be taken seriously. Unsustainable tourism development can result in environmental damage, exploitation of natural resources, socioeconomic inequality, and a decrease in the quality of life of local communities. A holistic and integrated approach in sustainable tourism development is needed to overcome this challenge. In addition, the principle of Bela Negara has excellent potential to be integrated into sustainable tourism development strategies in Sleman Regency. The focus of State Defense refers to the responsibility of every citizen to the interests of the nation and State, both in terms of defence and security as well as social and economic development. By integrating the principles of Bela Negara into the sustainable tourism development strategy, synergy can be created between sustainable tourism development and national development as a whole. (Agussalim, 2022) Copyright <unk> 2023 IJMSSSR All rights reserved However, despite the great potential and clear relevance, there is still a lack of understanding and application of the principle of State Defense in the context of tourism development in Sleman Regency. This lack of knowledge can hinder sustainable tourism development that pays attention to environmental, social, and cultural aspects and reduces the economic benefits that local communities can obtain. Therefore, research that integrates the principles of Bela Negara into sustainable tourism development strategies in Sleman Regency is significant. Thus, this scientific article presents theoretical foundations and practical recommendations regarding integrating Bela Negara principles in sustainable tourism development in Sleman Regency. Based on the background above, this scientific article will try to answer several main questions as a problem formulation: How is Bela Negara's principle integrated into the sustainable tourism development strategy in Sleman Regency? What is the positive impact of integrating Bela Negara principles and values in sustainable tourism development in Sleman Regency? The purpose of this article is first to Analyse the concepts and values of the principles of Bela Negara and understand their relevance in sustainable tourism development. Second, Explain the current State of tourism in Sleman Regency and identify the need to integrate the principles of Bela Negara in the sustainable tourism development strategy, as well as what is the positive impact of the integration of the principles of Bela Negara in sustainable tourism development in Sleman Regency. --- METHOD This research uses qualitative methods -descriptive and data collection through observation, interviews, and literature studies. (Creswell, 2013) The collection of references is mainly related to the concept of Bela Negara, the tourism potential in Sleman Regency, and the relationship between Bela Negara and sustainable tourism development. Data from literature sources is mainly obtained through searching for information from reliable sources such as journals, books, government reports, and related documents related to tourism and State Defense. The collected data was analysed using a descriptive approach to illustrate how the principles of Bela Negara are integrated into sustainable tourism development strategies and what positive impact it has on incorporating the principles and values of Bela Negara in sustainable tourism development in Sleman Regency. This article will be arranged systematically using a clear structure, such as introduction, background, purpose of writing, writing method, analysis results, and conclusion. Each section will be developed with supporting arguments and comprehensive explanations. Using descriptive-analytical research methods and qualitative approaches, this article is hoped to provide a complete understanding and in-depth analysis of the relationship between Bela Negara and tourism development in Sleman Regency DIY. --- RESULTS AND DISCUSSION: --- A. Basic Values of State Defense In conceptual analysis, it was found that the principle of Bela Negara includes values such as love for the motherland, loyalty, discipline, active participation, and responsibility to the interests of the nation and State. Meanwhile, Sustainable tourism development aims to achieve economic, social, and environmental sustainability in developing the tourism sector. Integrating the Bela Negara principle in the sustainable tourism development strategy can strengthen the contribution of tourism to national development and ensure that the interests of the nation and State are safeguarded. Bela Negara is a broad and holistic concept that includes the responsibility of every citizen in maintaining, defending, and developing the country. Bela Negara is not only related to national defence and security but also includes social, economic, political, cultural, and environmental aspects. In essence, Bela Negara involves actively participating in all components of society to advance the nation and State. (Ariyanto, 2013) In general, State Defense can be interpreted as concrete attitudes and actions taken by citizens to protect, fight, and build the integrity, sovereignty, and progress of the State. The concept of Bela Negara encourages active community involvement in various fields, including national development, maintenance of security and order, Copyright <unk> 2023 IJMSSSR All rights reserved human resource development, cultural preservation, and environmental management. In the context of tourism development, Bela Negara has a deep meaning. In addition to maintaining security and order in tourism destinations, Bela Negara also plays a role in preserving nature, preserving local culture, and building community awareness and responsibility towards the surrounding environment. This concept encourages active community participation in sustainable and responsible tourism management. (Kusumaningsih & Supriatnoko, 2022) Bela Negara is understood as a concept involving all elements of society, including the government, local communities, tourism actors, and tourists, to develop sustainable tourism in Sleman Regency DIY. Through this concept, it is hoped that tourism can become an instrument that encourages economic growth, cultural preservation, and nature preservation, as well as strengthening the sense of nationality and local identity. --- B. Sustainable Tourism According to Sharpley, the primary goal of sustainable tourism development is to balance the tourism environment with the needs of local communities and tourists. In other words, the purposes of achieving sustainable development are: a. Development goals focus on economic growth through a grassroots approach to development that focuses on satisfying the community's basic needs. b. Environmental/sustainable goals include conserving and protecting the environment, especially nonrenewable resources. (Yohanes et al., 2017) Müller explained that sustainable tourism aims to realise a healthy economy and a prosperous society, preserve local culture, and give optimal satisfaction to the needs of tourists. On the other hand, Hunter affirms sustainable tourism as "an adaptive paradigm that legitimises different approaches according to specific circumstances". According to Niedziolka, sustainable tourism is "all forms of tourism activity, management, and development that preserve natural, economic, and social integrity and guarantee the maintenance of natural and cultural resources".(Kisi, 2019) Sustainable tourism will make people responsible for respecting and preserving economic, environmental, and socio-cultural balance. It can be said that the tourism sector is the sector that has the most economic impact. It's just that only a tiny part of the existing potential has become a tourist attraction, and even then, not all of it is appropriately managed. (Sadali et al., 2020). The tourism sector will contribute to the community's welfare if addressed adequately. Tourism not only contributes to improving welfare but also has an impact on improving the quality of life of the community as a whole. (Hartwell et al., 2018). Tourism development is not enough to sell destinations but to present tourism that gives birth to happiness, satisfaction, and welfare so that, in the end, it will be able to improve the quality of life of tourists. (Yu et al., 2018). There are many considerations when someone decides to visit a destination, including the friendliness of residents to visitors, distance/flying time to the destination, ease of communication between residents and visitors, ease of local transportation, health / medical facilities, infrastructure adequacy, hotels and restaurants of good standards. For this reason, commitment between local governments, the private sector, and local communities is needed. (Petrovi<unk> et al., 2018) For the tourism sector to improve people's welfare, optimising sustainable tourism is crucial, namely, tourism that fully considers the social, economic, and environmental impacts and how to address the needs of visitors, investors, and the local community. This is done in a way that sustainable tourism must ensure optimal use of environmental resources, respect the socio-cultural characteristics of local communities, and provide socioeconomic utility to stakeholders. (Kisi, 2019). Cooper explained that at least four essential components must be met in tourism development: attractions, accessibility, facilities and extra services. (Astuti & Noor, 2016). Attractions offer the interests of a tourism area, such as natural beauty, regional culture, and others. Accessibility is the ease of transportation access to destinations and local transportation in the tourism area. Facilities are the availability of accommodation in tourism areas, such as lodging places (hotels, homestays, hostels, etc.), restaurants, health facilities, souvenir sales places, entertainment venues, waste/waste processing sites, electricity, clean water, etc. Extra Services, meaning the existence of tourism organisations needed for tourist services, such as hospitality associations, tour guide associations, and others. --- C. Tourism Development Based on State Defense in Sleman Bela Negara is essential in developing sustainable tourism in Sleman Regency DIY. In this context, Bela Negara can be interpreted as the efforts of the community and government in protecting and preserving natural resources and culture, as well as regional security and stability, to support sustainable tourism development. ( This is expected to move all components in the destination area so that naturally, there will be a process of community empowerment by placing residents as subjects of para-tourism actors. Breksi Cliff Park tourism, located in the Prambanan hilly area, previously a stone mining area, can be said to be proof of success in combining two successful tourism concepts. Breksi Cliff Park has grown significantly to increase employment and improve the welfare of the local community. (Antaranews, 2022) In addition to Breksi Cliff Park, Sleman Regency has several other tourist destinations. At least it can be grouped into four categories, namely: a. Nature Tourism Sleman Regency DIY has stunning and diverse natural beauty. With its location at the foot of Mount Merapi and bordering the Indian Ocean, Sleman offers a spectacular landscape. Some of the potential natural beauty of Sleman Regency DIY include: 1. Mount Merapi is one of the main attractions of Sleman Regency. With the enchanting beauty of the peak, tourists can enjoy stunning natural panoramas and challenging trekking. Even Volcano Tour has become one of the favourite tourist destinations among tourists. 2. Kaliurang is also a mountainous area located on the slopes of Mount Merapi. This place offers cool air, beautiful scenery, and various activities such as tracking, horse riding, or visiting the volcanology museum. --- b. Historical and Cultural Heritage The Sleman Regency area is one of the districts with the most cultural heritage sites. Each sub-district in the Sleman Regency has archaeological sites of the nation's cultural heritage. In total, there are 181 archaeological sites in Sleman Regency. (Pradoko, 2017) Several historical and cultural heritage sites are spread across 17 sub-districts in Sleman Regency. There is not a single sub-district that does not have an archaeological site. With details: Gamping sub-district has as many as eight areas, Godean has seven sites, Moyudan This article has uncovered the significant benefits of integrating Bela Negara principles in sustainable tourism development in Sleman Regency through conceptual analysis, field research, and data analysis. These benefits include environmental and natural resource sustainability, economic and social empowerment of local communities, cultural preservation, and improving the image and attractiveness of Sleman Regency as a tourism destination. However, challenges such as sustainable environmental management, community participation, and cultural preservation were also identified. Therefore, practical recommendations are given to address these challenges, such as encouraging active community participation, integrating the principle of Bela Negara in tourism policy, developing cultural preservation programs, and enhancing cooperation between the government, tourism industry, and communities. Integrating Bela Negara's principles is essential to developing sustainable tourism in Sleman Regency and provides a solid moral foundation and national interest. By implementing these recommendations, sustainable tourism development in Sleman Regency is hoped to run harmoniously, provide sustainable economic, social, and environmental benefits, and strengthen local identity and pride.

This article discusses integrating State Defense principles and values into the sustainable tourism development strategy in Sleman Regency. Tourism is essential for Sleman Regency, but rapid growth brings environmental, social, and cultural challenges. With a diverse social structure regarding ethnicity, religion, and socio-economic level, Sleman Regency has the potential for friction between different groups. However, these groups have synergy and mutual tolerance in developing sustainable tourism in Sleman Regency. On that basis, this research aims to answer what factors bind them together. This research uses a qualitative-descriptive method, with data collection techniques by conducting observations, interviews, and literature studies. The finding of this research is that the main factor that causes diverse communities to synergise and tolerate each other in the development of sustainable tourism in Sleman Regency is the national awareness of the people in tourist destinations of the responsibility of citizens to the interests of the nation and State in development so that this attitude strengthens the sustainable tourism development program. This article concludes that Sleman Regency integrates the principles and values of State Defense in tourism policy, develops cultural preservation programs, and increases cooperation between the government, tourism industry, and the community as part of a sustainable tourism development strategy. Hopefully, this article can contribute thoughts and guidelines for policymakers and related parties in developing sustainable tourism that considers the principles of State Defense in Sleman Regency.

Introduction In healthcare settings, lesbian, gay, bisexual, transgender, and queer (LGBTQ+) populations often experience discriminatory environments resulting from explicit or implicit biases held by healthcare providers [1][2][3]. Current literature has identified a positive association between transgender patients' experience of discrimination in oral healthcare settings and their level of dental fear [4]. A study conducted by the Center for American Progress (2018) indicated that discrimination experienced by LGBTQ+ patients in healthcare settings may largely affect their ability and intent to access care [5]. Dental anxiety resulting from explicit or implicit biases in oral healthcare settings can prevent LGBTQ+ patients from seeking or following through with regular dental care appointments may lead to adverse oral health outcomes. Existing literature reveals disparities in overall health, including oral health, among the LGBTQ+ communities [6,7]. Lesbian, gay, and bisexual (LGB) individuals have a higher risk of poor mental health, smoking, and substance misuse than heterosexual individuals [8,9]. Increased risk of oral disease is found in the LGBTQ+ population because of depression, side effects of medications, and harmful eating behaviors [10]. Human Papillomavirus associated oropharyngeal cancers and Human Immunodeficiency Virus infections, among other sexually transmitted diseases with oral health implications, are also more prevalent in the LGBTQ + population [11]. Limited evidence from literature indicate there may be an association between hormone replacement therapy (HRT) and periodontal disease in transgender community [12,13]. Due to the associated stigma, many transitioning transgender patients refrain from discussing this with their oral health providers and may develop concerns as possible effects of HRT on oral health [14]. Oral health is often a mirror image of an individual's overall health; in fact, several oral diseases are associated with systemic conditions and vice versa. Periodontal disease negatively affects glycemic control; this relates to the causal relationship between poor oral health and diabetes [15,16] Oral health outcomes affect more than just the periodontium. In addition to diabetes, poor oral health has been linked to cardiovascular and rheumatoid disease [17][18][19]. Oral health is also known to have a positive relationship with self-esteem [20], positive quality of life and affects presenteeism, performance at work and academic achievement [21][22][23]. Although there is an upward trend in the acceptance and visibility of transgender and nonconforming adolescents in seeking oral health services, a great deal of work still needs to be done to improve their patient experience while navigating the system [24] and reducing disparities in accessing oral health services. A recent study conducted across different disciplines, specifically reported that dental students had significantly less positive perception of their formal training in LGBTQ+ health and more stereotypical attitudes towards LGBTQ+ populations, when compared to medical and nursing trainees [25]. With this project, we aim to evaluate the experiences of LGBTQ+ patients in oral healthcare settings including their health seeking behaviors alongside the beliefs, perceptions, and attitudes of oral healthcare professionals toward the LGBTQ+ community and towards implementing inclusive oral healthcare practices in their offices. --- Methods --- Data source and ethical considerations From August 2020 to February 2021, descriptive, quantitative surveys were administered to OHPs and LGBTQ+ patients. For this study, OHPs were defined as dentists, dental hygienists, dental assistants, and administrative support staff (in dental clinics) and LGBTQ+ patients were defined as individuals who were part of these clinics' patient base and who self-identified as being either a sexual and/or gender minority. The Indiana University School of Dentistry Institutional Review Board approved this investigation (Protocol #: 2007467343) as an exempt study because the study did not require the collection of any identifiable data from any participants. This designation waived the IRB need for formal written consent. Respondents read a statement online which informed them that the study was completely voluntary, summarized the purpose of the study, and how their data would be used. Respondents were informed that by completing the survey they were offering their consent to participate. All survey data were stored in the Qualtrics online data application (26). The authors disclose no conflicts of interests. The participants (both OHPs and LGBTQ+ patients) were recruited from dental clinics at Indiana University School of Dentistry, invited federally qualified health centers (FQHCs) and community clinics within the states of Indiana and Michigan. Electronic survey links were shared with the oral health providers' listservs at the collaborating organizations. The survey for patients was distributed via the patient portal at these clinics and included a screening question which asked participants of their gender orientation and they could proceed with the survey only if they confirmed they were part of the LGBTQ+ community. Participation for LGBTQ+ patients were incentivized and patients who completed the survey were provided a $10 Amazon gift card. There we no incentives offered to the OHPs. --- Survey tool Both surveys contained multiple choice questions about participant demographics, including their gender and sexual minority status. The survey administered to OHPs consisted of 29 questions and included Likert scale questions about OHP's perceptions of the LGBTQ+ community, presence of inclusive healthcare practices within their office settings and their attitudes toward implementing changes to improve patients' quality of care in terms of inclusive healthcare practices. The term "inclusive healthcare practices" referred to questions around medical history, office environment and use of preferred pronouns for the LGBTQ+ patients. The survey administered to LGBTQ+ patients was 21 questions long and included Likert scale questions on their oral healthcare seeking behaviors, perceived barriers to accessing oral healthcare, and experiences relating to accessing oral healthcare. Both surveys are attached under Appendix I in S1 File. --- Dependent variables The outcomes of interest for this study were LGBTQ+ patients' comfort in attending dental appointments and OHP's sense of responsibility to treat patients who are members of the LGBTQ+ community. LGBTQ+ patient comfort in attending dental appointments was measured by a single Likert scaled item asking patients whether "[they are] comfortable going dental appointments." This item was dichotomized to compare strongly disagree/disagree/and neutral relative to responses of agree/strongly agree. OHP sense of responsibility to treat members of the LGBTQ+ community was similarly measured by a single Likert scaled item asking OHPs whether "[They] feel it is [their] responsibility to care for patients who are part of the LGBTQ+ community." This item was also dichotomized to compare strongly disagree/disagree/and neutral relative to responses of agree/strongly agree. --- Independent variables LGBTQ+ patient perceptions regarding the presence of inclusive healthcare practices in the oral healthcare setting where they receive care were used to predict LGBTQ+ patient's comfort in attending dental appointments. Inclusive healthcare practices included the presence of posters or artwork that specifically caters to the LGBTQ+ community, the ability of patients to indicate their preferred pronouns on medical forms and being treated the same as patients who are not members of the LGBTQ+ community. To predict OHP's sense of responsibility to treat members of the LGBTQ+ community, questions related to OHP self-reported affiliations (with the LGBTQ+ community) were used. --- Statistical analysis We used counts and percentages to describe the population characteristics of study participants and to explore the distribution of participant responses to measures of inclusive healthcare practices in oral healthcare settings. LGBTQ+ status among both participant groups were also compared descriptively. Subsequently, logistic regression models were constructed to predict LGBTQ+ patient's comfort attending dental appointments and OHP's sense of responsibility toward treating LGBTQ+ patients. Each model's design was informed by the previously described independent and dependent variables. These models were evaluated at a P <unk> 0.10 significance level and did not include demographic confounders. This exclusion was made to avoid bias related to this study's use of convenience sampling and the class imbalance caused by participant self-reporting. All analyses were completed using an opensource statistical software, RStudio. --- Results In total, 255 OHPs and 248 LGBTQ+ patients participated in this study. Of enrolled LGBTQ + patients, the median age was 29 years (IQR 25-36). This sample was demographically diverse with 89% (n = 221) of LGBTQ+ patients identifying with a non-heterosexual sexual identity and 21% (n = 52) identifying with a transgender gender identity. Among enrolled OHPs, the median age was 46 years (IQR, 33-59), with 63% (n = 145) reporting having worked in their current setting > 4 years. Within the OHP population, 52% (n = 132) were dentists, 14% (n = 36) were dental assistants, 12% (n = 31) were dental hygienists, and 20% (n = 51) were administrative staff members. Demographically, we observed that our population of OHPs was much more hegemonic with 89% (n = 206) of OHPs identifying as straight or heterosexual and 100% (n = 231) identifying as cisgender. When OHPs were asked to self-identify and define their specific affiliation with the LGBTQ+ community, we found that 28% (n = 66) had no affiliation, 52% (n = 120) had a family member or close friend who is a member of the LGBTQ+ community, 28% (n = 64) were allies to the LGBTQ+ community, and that only 5% (n = 11) self-identified as a member of the LGBTQ+ community. --- Participant characteristics For the LQBTQ+ patients enrolled in this study, we observed high levels of oral healthcare seeking attitude with 75% (n = 187) believing that dental care is a necessity and that it affects one's overall health. This additionally translated into a moderately high level of oral healthcare maintenance with 71% (n = 168) of LGBTQ+ patients reporting to have regularly (at least twice a year) visited their dentist prior to the onset of the COVID-19 pandemic. In contrast to these levels, we found that 43% (n = 106) of LGBTQ+ patients felt uncomfortable seeking care and that 34% (n = 84) believed they had been treated unfairly in oral healthcare settings because of their sexual identity or orientation (Table 1). --- Patient oral healthcare seeking behaviors and perceptions of oral healthcare settings (Table 2) Through logistic regression modeling we identified that, when present in oral healthcare settings, inclusive healthcare environments in oral healthcare settings were found to predict significantly higher odds of patient comfort (Model 1, P<unk>0.10). Of all the inclusive healthcare strategies, the ability for patients to indicate their preferred pronouns on medical history When investigating the barriers faced by our LGBTQ+ patient population, we found that financial barriers to care affected 50% (n = 124) of our patient population. While patients also reported barriers in scheduling appointments (Table 2.), this impacted a minority of participating patients. However, we ultimately observed that the majority of LGBTQ+ patients (64%, n = 159) believed that people in the LGBTQ+ community have additional obstacles accessing healthcare. --- Model 1: Environment based predictors of patient comfort in oral healthcare settings (Table 3) Overall, a large majority (84%, n = 191) of OHPs believed that the oral healthcare setting where they work has created a welcoming environment for members of the LGBTQ + community and 98% (n = 223) believed that the staffs there treat patients who are members of the LGBTQ+ community the same as those who are not. More than half 54% (n = 123) reported positively when asked whether they knew who they should ask if they had a question relating to LGBTQ+ care. A difference in beliefs, attitudes, and perceptions of the LGBTQ+ community was observed across all four OHP groups (Fig 1). While these differences were not statistically significant, we observed that, on average, dentists were less likely to agree that they were comfortable providing care to LGBTQ+ patients, were comfortable working with LGBTQ+ coworkers, reported feeling they had a responsibility to care for patients who are members of the LGBTQ+ community, and were willing to learn and utilizing resources to improve access to care for patients who are a part of the LGBTQ+ community. Further, across all provider groups, we observed low levels of agreement relating to OHPs being aware of the health disparities faced by LGBTQ+ populations, having undergone adequate training regarding cultural competencies and the LGBTQ+ community, and being aware of the impacts of HRT on an individual's overall health and specifically on oral health. --- Model 2: Affiliation-based predictors of perceived responsibility to treat LGBTQ+ patients among dentists (Table 4) Among dentists, we explored how affiliation to the LGBTQ+ community might impact a dentists' sense of responsibility to treat members of the LGBTQ+ community. While there was no statistically significant relationship found between dentists identifying as having no affiliation or as being a member of the LGBTQ+ community and their sense of responsibility, we did find that either having a family member or close friend who is a member of the LGBTQ+ community or identifying as an ally produced a statistically significant odds (P<unk>0.01) of feeling a responsibility to treat LGBTQ+ patients. --- Discussion In our sample of LGBTQ+ patients and OHPs, we observed a need to improve oral healthcare practices in dental settings and to provide educational trainings for OHPs related to LGBTQ + populations and their health. Our results were comparable to some of the previously conducted research in terms of difficulties in accessing care among the LGBTQ+ patients and the need for a holistic and welcoming attitude among health professionals [26]. This may be different from few of the other studies which report transgender and non-conforming adolescents and young adults have minimal difficulty receiving oral health care [24,27]. Although there may not be much "tooth level" difference in treating this population, there is an obvious need for developing trusting, non-judgmental patient-doctor relationship in caring for this vulnerable group [28]. Literature indicates that knowledge and attitudes among healthcare providers about homosexuality, are influenced by them being an ally or having a friend who is from the LGBTQ+ community [29]. Increasingly over the past decade, public health literature has emphasized the role that inclusive healthcare environments play in improving health outcomes among marginalized populations (2). Related interventions often take the form of educational toolkits or continued education opportunities; however, in some instances, interventions have also been deployed at schools of medicine in the form of curricula changes [25]. Within the field of public health dentistry, there has been little to no emphasis on the development of such interventions to improve care of marginalized populations or, more specifically, LGBTQ+ patients. Our results indicated that introducing established, inclusive healthcare practices would improve the comfort of LGBTQ+ patients in oral healthcare settings. We found that, while agreement differed across provider types, all groups expressed moderately high agreement that they were willing to learn and utilize resources to improve access to care for LGBTQ+ patients. This level of agreement indicates a similarly high level of interest among OHPs to engage with LGBTQ + health education and training. --- Limitations Our study was limited by several factors. For instance, the convenience-sampling techniques limit the representativeness of the sample. Our patient population included those that were currently receiving dental care, that does present a selection bias. Many LGBTQ+ patients, delay or completely forgo medically necessary care for fear of stigma unlike our patient sample that had a dental home. This study was a cross-sectional, exploratory study due to which, we are unable to make claims of generalizability of the study results. Additionally, we did not formally validate our survey prior to data collection, though items were reviewed by an interprofessional group of collaborators where the study was implemented. With the topic in question, social desirability bias may be another limitation, especially for the OHP's survey. Although the surveys were anonymous, many participants may not have responded honestly about negative attitudes towards or discomfort with LGBTQ individuals. It's important to note that information on the educational status of the patient pool which is an important variable and can directly affect their health literacy and ability as well as intent to avail oral healthcare services, wasn't collected. The study sites were in relatively conservative states of Indiana and Michigan which may as well confound the study results. There are currently no comprehensive civil rights protections in place for LGBTQ individuals at the state level in Indiana [29]. Michigan has one of the highest levels of hate crimes reported per 100,000 residents and the overall environment is not very LGBTQ+ friendly [29]. --- Further research and practical implications The need to improve the health, safety, and well-being of LGBTQ+ individuals is also listed as a goal for Healthy People 2020 [30]. Poor attitude towards LGBTQ+ individuals can affect resource utilizations and LGBTQs health status implications for oral health [31]. Future research in this direction will help identify gaps and direct oral health policies to improve access to care for the LGBTQ+ community. Provider stigma attached with treating this population is a huge issue and motivating providers to engage with this population and influencing their attitudes may require a lot of efforts in terms of cultural humility and workplace inclusiveness training for the OHPs. Educators should develop training programs that provide OHPs with the knowledge and skills to ensure LGBTQ+ patients receive effective oral health care when they access services for themselves. --- Supporting information S1 File. This is the S1 File (Appendix). This is the S1 Appendix I.

In healthcare settings, lesbian, gay, bisexual, transgender, and queer (LGBTQ+) populations often experience discrimination, leading to decreased healthcare services utilization. In this study we have tried to identify oral healthcare providers (OHP)'s perceptions toward LGBTQ+ patients, perceived barriers for LGBTQ+ patients in accessing oral health services, and whether they were open to inclusive oral healthcare practices. In addition, the experiences of LGBTQ+ patients in oral healthcare settings including their oral healthcare seeking behaviors and beliefs were also explored.Descriptive, quantitative surveys were administered to OHPs and LGBTQ+ patients within Indiana and Michigan. Surveys contained questions about participant demographics, including gender and sexual minority status, and the presence of inclusive healthcare practices within the oral healthcare settings. Descriptive analyses and regression modeling were used to explore the distribution of participant responses and to identify predictors associated with patient comfort and OHP's attitudes toward LGBTQ+ patients.Overall, 71% of LGBTQ+ patients reported regularly attending dental appointments; however, 43% reported feeling uncomfortable going to appointments and 34% reported being treated unfairly during appointments because of sexual orientation. Among OHPs, 84% reported that the healthcare settings where they practiced were welcoming for LGBTQ+ populations and 84% reported willingness to improve LGBTQ+ care. The presence of inclusive healthcare practices predicted comfort for LGBTQ+ patients (P < 0.10). Additionally,

Introduction Maternal health care (MHC) utilization is essential for women's health, childbirth, and the well-being of the mother and child. Maternal health care includes the care a woman receives throughout her pregnancy, labor, and postnatal [1]. The World Health Organisation has set a goal in alignment with the 3rd Sustainable Development Goals (SDG) which aims to reduce the global maternal mortality ratio (MMR) to less than 70 per 100,000 live births by 2030′′ [2]. To achieve this goal, ascertaining the determining factors of maternal health care utilization and how these factors affect the livelihood of mothers is critical and must be a priority in policy development [3]. Globally, 810 women die per day owing to pregnancyrelated complications [4]. Maternal death remains a great concern with virtually 99% of all maternal deaths occurring in developing nations, but the prevalence is higher in sub-Saharan Africa [5]. Sierra Leone has one of the world's highest maternal death ratios at 1360 deaths per 100,000 babies born. WHO estimated that up to 6% of women in Sierra Leone will die from maternal causes during their reproductive life" [2]. Research also suggests that this high prevalence of maternal, neonatal, and child death are linked to non-or poor availability of quality maternal healthcare services [6,7]. Again, in sub-Saharan Africa, the factors that influence the increase in maternal mortality are also associated with prenatal care coverage, skilled attendance at delivery and postnatal care. The non-use of these maternal healthcare services is a key predictor of perinatal mortality [2]. Accessing skilled care before, throughout pregnancy, and after delivery can therefore save the lives of mothers and their newborn child infants [7,8]. If mothers have access to healthcare facilities that provide interventions and preventive measures to treat obstetric complications, especially in an emergency, then an estimated 74% of maternal mortality could be prevented [7]. Antenatal care, skilled delivery attendants, and postnatal care attendants therefore remain key providers to improve health outcomes for mothers and their babies [9]. Studies in Ethiopia, Ghana, and Cambodia showed similar findings [6,10,11]. Uptake of these MCH services enables the healthcare providers to have a better chance to detect and reduce any risk factors associated with adverse pregnancy outcomes and also serves as a platform to counsel the mothers on why a skilled attendant is crucial at delivery [12,13]. Though the use of maternal health care services offers better health outcomes for mothers and their children [14], only 45% of women using skilled delivery attendants at delivery in sub-Saharan Africa [5] and up to 13% of women in Sierra Leone still using unskilled delivery attendants [11]. There is the need to explore factors that influence the use of MCH services such as the effect of demographic factors on service utilization in Sierra Leone. Effects of these factors remains unclear, especially in districts far from the national capital such as Kailahun district, the epicenter of the 2014 Ebola outbreak. Again, the Sierra Leone Demographic and Health Survey (SLDHS) covers some components of MHC service, but details on specific factors influencing utilization are not fully addressed. Again, other previous studies conducted in Sierra Leone focused on one or two of them [15,16]. This research sought to investigate the socio-economic factors that influence women's use of maternal healthcare services in Eastern Sierra Leone with a focus on ANC, delivery care, and postnatal care. It will provide a further understanding of factors influencing uptake of maternal health services in Kailahun, Sierra Leone, thus helping in implementing policies that would improve maternal and child health. --- Methods --- Study design, population, and setting We conducted a community-based cross-sectional study that involved women of reproductive age (15-49 years), who had at least one delivery in the 3 years before this study in the Kailahun District, Eastern Sierra Leone from November 2019 to January 2020. Kailahun District is located about 450 km from Freetown (the capital) in the Eastern province of Sierra Leone. Its capital and largest city is the town of Kailahun. It is divided into 15 chiefdoms. Healthcare facilities are divided into 15 community health centers (CHC), 52 community health posts (CHP), 17 maternal child health posts (MCHP), 3 government hospitals, and 2 private clinics. The total projected population for 2020 is 625,500 [17]. The total fertility rate is estimated at 6 children per woman [17]. The ANC attendance for ANC1 and ANC4+ was 72.5 and 63.9% respectively [14]. The District of Kailahun shares borders with the Republic of Liberia to the East, and the Republic of Guinea to the North. --- Sampling method The sample size was determined using Cochran's formula [15] with the following assumptions: the estimated proportion (p) of births by SBA was 0.6, the estimated proportion of births by an unskilled attendant (q) was 0.4, and a design effect (DEFF) of 1.5 from a previous study conducted in Sierra Leone [16]. The level of precision (d) was 0.05, and a 95% CI (z) was 1.96, were used to obtain a sample size of 554. For sampling of households, the 2015 census EAs were used as the Primary Sampling Unit (PSU) with the Urban and Rural areas of Kailahun District as the domains. The proportional allocation of household in the Kailahun n = z 2 pq d 2 X DEFF District based on the domain was 69% rural and 31% urban. A Random-walk technique was used in selecting households/respondents. At the center of the EA, the interviewer identified the third house from his righthand side. Upon arrival in the HH only one eligible female 15-49 years had given birth in the past 3 years before this study was enrolled. Where there was more than one woman eligble, women balloted by picking yes or no on small sheets of paper. The interviewer moved in the eastward direction into every third house. If no one is eligible for a particular HH, we continued with the pattern of selecting the HH. --- Data collection technique and tools The primary data was collected using a structured and pre-tested questionnaire. Antenatal cards and hospital record books of mothers were inspected to confirm the information provided by women during the interview. The questionnaire was written in English and was adapted from the Standard Sierra Leone DHS Questionnaires [16]. --- Data quality management To ensure quality data, interviewers were trained, a pretest was performed before the actual data collection. A supervisor was assigned to interviewers throughout the data collection period to ensure data was collected appproriately. Daily spots checks of filled questionnnaires was done and issues itendified resolved before the day's work began. Data validation checks were done to ensure accuracy during data entry. --- Operational definitions • Skilled delivery attendant: professionals with midwifery skills including Doctors, midwives, community health officers (CHO). • Antenatal care: Care provided by skilled healthcare professionals to pregnant women and adolescent girls in other to ensure the best health conditions for both mother and baby during pregnancy. • Postnatal care: services provided to women within 42 days after delivery by health professionals. --- Outcome variables A set of questions on maternal healthcare, covering place of delivery, delivery attendant, place of PNC, PNC attendant, the timing of the first postnatal visit, follow-up after discharge, and place of follow-up after discharge were asked. For the study, the outcome variable which was the utilization of maternal health care services was assessed by two services: --- Independent variables Andersen's behavioral model was adopted to determine the influence or effect of the independent variables (predisposing and enabling factors) on the utilization of maternal health services [18,19]. The predisposing factors were age, women's education, parity, ethnicity, religion, marital status, occupation, number of births in the last 5 years, number of children alive, and the enabling factors were husbands education, residence, distance to the health facility, type of floor and toilet. --- Data analysis The data were entered into MS Excel 2018 and cleaned. The analysis was done using Stata IC 15.0, College Station, TX: StataCorp LLC. Frequencies and proportion were generated to describe the Predisposing and Enabling characteristics, the components/patterns of antenatal care, perinatal care/skilled delivery attendant, and postnatal care. Frequencies were presented as tables. Bivariate analysis was done to determine the level of utilization of Maternal Healthcare services (Antenatal care, Skilled delivery attendant, and Postnatal care services). In the case of a skilled delivery attendant, a dichotomous dependent variable was constructed to indicate whether or not the woman used services from a skilled provider. Bivariate and multivariate analysis techniques were used to estimate the nature of association and magnitude between dependent variables and independent variables (Individual and Enabling factors). Initially, a simple binary logistic model was run for each independent variable against the outcome variable. All independent variables that showed a significant association at p <unk> 0.05 were included in the multivariate binary regression model. The P-value, crude odds ratios (COR), adjusted odds ratios (AOR), and 95% confidence interval were estimated. --- Ethical clearance Ethical approval was obtained from the Sierra Leone Ethics and Scientific Review Committee (October 2019). Permission was also sought from the Kailahun Regional Health Directorate. Informed consent was obtained from the respondents. For respondents below 18 years, consent was granted by the parent or guardian, and assent was sought from respondents. They were assured of confidentiality and informed of the purpose of the study. The information was stored without the names of the respondents in a folder that is only accessible to the data team. For participants who could not read, the consent was read and explained to them in the presence of an impartial witness or stakeholder in the community who also appended his signature on the consent form in accordance with the local IRB regulation. All study methods were carried out following relevant guidelines and regulations. --- Results --- Background characteristics of respondents Five hundred and fifty-four (554) women aged between 15 and 49 years, who had at least a delivery/birth 3 years before this study commenced were interviewed. The median age of respondents was 25 years with minimum and maximum ages of 15 and 49 years respectively (Table 1). Taking into consideration parity, the median parity was 3 with a minimum and maximum parity of 1 and 9 respectively. Forty-six percent (46%) of respondents in this survey were between the ages of 25-35 years and 37.9% had no educational level. The majority (89.5%) of respondents were either married or living together with a partner, 6.7% were single and 3.8% were divorced/ Separated/widowed. Most of the respondents (68.8%) were rural residents. The majority (60.8%) of respondents were Muslims. The Mende ethnic group was predominant, (72%). --- The components of ANC, skilled delivery, and PNC utilization The 2). --- Socio-demographic characteristics and the use of MHC The level of utilization of skilled ANC attendants was more common among respondents aged > 35 years compared to the other age groups. In general, the use of skilled ANC attendants was 75.0% or more among all background characteristics assessed. With regards to parity, as the parity increased the use of skilled ANC attendants decreased, more women with no education attainment received at least one ANC visit (78.6%) compared to educated women (77.5%). The use of SDA was found to be higher (48.7%) amongst single compared with married women 34.7% (62/172). Similar to ANC use, the use of skilled birth attendants decreased as the parity increased, 43.2% (57/132) of women who had 1 parity used skilled birth attendants compared with 37.3% (226/292) of those who had 2-4 parities and 25.4% of those who had 5 or more births. More than 90% of all women interviewed received PNC services. All women whose husbands were unemployed received PNC services. In contrast to ANC and skilled birth attendant use, PNC service use was slightly higher (97.9%; 372/381) among women in rural areas than women in urban areas (96.5%; 167/173) (Table 3). --- Background characteristics influencing the use of a skilled ANC attendant Maternal age was not significantly linked with the use of skilled ANC providers. Women aged > 35 years were more likely to use a skilled ANC attendant than women aged 10-19 years which was found not to be significant (COR 1.71, 95% CI 0.68-4.28). There was no significant association between marital status and the use of a skilled ANC provider. The odds of a skilled ANC attendant use were 1.66 times higher in divorced/separated/widowed compared to single women which were not statistically significant (COR 1.71, 95% CI 0.68-4.28). There was no significant association between religion, ethnicity, occupation, education, and the use of skilled ANC providers (Table 4). --- Background characteristics influencing the use of a skilled delivery attendant Place of residence, parity, occupation, educational attainment, husband educational attainment, distance to health facility were significantly associated to skilled delivery attendant (Table 5). Women residing in urban areas had over six-fold increased odds of being delivered by a skilled attendant as compared to women residing in rural areas. This association remained statistically significant even after controlling for the effect of the other variables (AOR = 6.20, 95% CI = 3.61-10.63). Women who had a senior secondary or vocational level of education had 1.22 times increased odds of being delivered by a skilled attendant as compared to those with no educational level which was statistically not significant after controlling for the effect of other variables (AOR = 1.22, 95% CI = 0.57-2.59). --- Background characteristics influencing the use of postnatal care services The level of education of a woman, residence, and the husband's education level, were significantly associated with the use of the standard PNC category relative to the average category after controlling for the effect of the other variables in the model ( --- Discussion This study assessed the level of utilization and determinants of ANC, skilled birth attendant, and uptake of the different packages of PNC services in Kailahun District Sierra Leone. This study found that 100.0% of women received ANC services, of which 77.6% of them sought at least one ANC visit from a skilled ANC provider and 88.6% made 4 or more ANC visits as recommended by the World Health Organization. These findings are similar to reports in the Sierra Leone Demographic and Health Survey which showed that 98.0% of women received ANC from a skilled provider and 79.0% made 4 or more ANC visits [11]. However, our results are also inconsistent with the Uganda Demographic and Health Survey UDHS 2016, which showed that 97.0% of women received ANC from a skilled provider but only 60.0% of the women made 4 or more ANC visits [20]. The disparity between the present study and the UDHS may be related to the maternal demographic characteristics in both countries and the fact that the sample size in the UDHS was far larger, thus influencing the precision of the findings. Regarding the timing of the first ANC, 54.5% of women sought their first ANC in the first trimester of pregnancy and this prevalence was higher than that obtained in the Sierra Leone Demographic and Health Survey 2019, where 44.0% made their first visit in the first trimester [11]. This discrepancy could be attributed to the demographic survey having had a larger coverage area as compared to the present study. It has been recommended that all pregnant women should start their ANC in the first trimester [21]. The findings of this research suggest that the use of ANC services was higher among urban residents (81.5%), which corroborates a study conducted in Holeta Town, Ethiopia where 86.7% of urban women used ANC services [22]. According to the Sierra Leone Demographic Survey (2019), 73.0% of women in urban areas made 4 or more ANC visits which is slightly lower patronage as compared to our population. The high use of ANC services in urban areas may be because, Kailahun District has 87 peripheral health units and 3 hospitals, of which all the hospitals and most of the PHUs are in urban areas thus increasing access. Increased awareness and information sharing might also be related to the economic status of urban residents since women in urban areas have more physical and economic access to health facilities. The findings of this study showed that the enabling (maternal age, marital status, ethnicity, parity, respondent education, and religion) and predisposing factors (residence, distance to the health facility, and husband's education) were non-predictor of ANC service utilization. A study conducted in Holeta Town, Central Ethiopia found maternal age and education as predictors of ANC service utilization [22]. In another study in Ghana, residence and education were major predictors of ANC services [1,23,24]. This research further indicated that the use of SDA was generally low. Only 35.9% of women had at least one delivery in 3 years before this study used SDA compared to 11.7% in a study in Ethiopia [5]. The majority, 57.6, and 4.2% were delivered by Maternal and Child Health Aide (MCHA) and Traditional Birth Attendant (TBA) respectively, who are considered non-skilled. According to the Demographic and Health Survey (2019), 87% of deliveries were assisted by a skilled provider. The disparity between the DHS and this study may be related to the fact that our study did not consider MCHA and State Enrolled Community Health Nurses (SECHN) as skilled attendants. Kailahun District is one of the most remote areas in Sierra Leone, physical accessibility is a major challenge due to the bad road network, thus most healthcare workers find it difficult to travel, work and stay in the district which creates opportunities for non-skilled workers. The WHO has recommended that there should be a critical threshold of 23 skilled healthcare workers (doctors, nurses, and midwives) per 10,000 population [25]. Nevertheless, it has been very difficult for Sierra Leone to cope with such recommendations due to the severe scarcity of qualified healthcare workers, thereby providing merely 2 skilled workers per 10,000 populations [14]. The 10 years of civil war which ended in 2002 and the 2014 Ebola epidemic, all started in Kailahun District and left a huge impact on health service delivery in the district. Sierra Leone is among the world's highest maternal death ratios at 1360 mortality per 100,000 births [2] because most women are not delivered by a SDA, and most ANC services are provided by non-skilled providers. It further found that 10.5 and 35.9% of deliveries took place at Maternal and Child Health Post (MCHP) and Community Health Post (CHP) respectively, which are facilities manned by non-skilled attendants. This study found that the area of residence is a major determinant of SDA utilization. The use of skilled birth attendants was higher among urban residents than rural. These findings are consistent with other studies [5,7]. The disparity in the utilization of MHC services may be due to the concentration of health facilities in urban areas combined with the high number of qualified birth attendants in urban areas and also the economic status of the urban residents. In the Kailahun district, there is an uneven distribution of health workers, most are found in urban areas. Distance to the health facility was considerably connected with the use of SDA. Women that walked 30-60 min or more than *Significant association at P <unk> 0.05 ** Variables adjusted for the effect of the other variables 60 min to access health care services were more likely to use SDA than those that walked less than 30 min. In another study conducted in Kenya, although the distance was cited as a barrier to MHC service utilization, 18% of women did not visit the nearest facility [26]. In the present study, the husband's education was significantly associated with the use of a skilled delivery attendant. Women whose husbands had at least primary education were more inclined to use SDA than those whose husbands had no education which is consistent with other studies [5]. Similarly, in another study conducted in Nigeria, husbands' education played a key role in the utilization of SDA [1], research has shown that education increases health awareness and knowledge on the significance of MHC services and improves other forms of learning [5]. This could be through; radio, the internet, written information, and a better cultural understanding. Educated husbands may provide more autonomy to their wives [5]. Our study found that the majority 97.5% of women received PNC services and 58.1% of them received it from MCHA. The 2013 Sierra Leone Demographic and Health Survey reported that 7.8% of PNC services were delivered by MCHA [11]. The disparity between the national and district figures may be related to the fact that the Demographic Survey was done in the entire 16-district taking into consideration the major urban areas where good healthcare services are concentrated compared with Kailahun being one of the most remote districts. Also, due to the limited number of skilled providers in the district, most PNC services are provided by non-skilled providers. In the present study, women's residence, education of women, and husband's education are significant predictors of the utilization of the standard PNC package. These observations are consistent with a study conducted in Ethiopia among women of reproductive age which reported a significant influence of respondent education and urban residence on the utilization of MHC services [5]. Some of the results of this study are similar to others reported, which have highlighted the importance of effective, efficient, and accessible Maternal healthcare services. Several issues specific to Sierra Leone were identified that suggest the influence of these socioeconomic factors on the utilization of MHC services and if these are not addressed in addition to the inadequate number of skilled attendants, Sierra Leone would continue to have high maternal and infant mortality ratios. Finally, our results show that investment in the training of skilled attendants, education, and rural healthcare would significantly improve the utilization of MHC service thus reducing maternal and infant mortality. --- Limitation of study Some of the respondents seemed to have difficulty in the recollection of events that had happened during the last 3 years before the study. Such women had difficulties in recalling or identifying the nature of healthcare services they received or the trained healthcare worker that provided the service. As a way of minimizing this challenge, the interviewers requested for participants' ANC cards to fact-check their responses. This study is externally valid and generalizable as the sample accurately represents the population and the characteristic of Kailahun District are similar to others in Sierra Leone. The households were proportionately distributed to urban and rural areas. We did not assess the healthcare facility and the health workers' related factors that affect antenatal, postnatal, and skilled delivery utilization, all of which are known to be key determinants of maternal health care utilization. These are areas for further study. --- Conclusion Our findings indicate that PNC service utilization in the Kailahun district is much higher compared to skilled ANC and birth attendant utilization, even though the uptake of the standard PNC package was low. The utilization of SDA was low and urban residents seemed to be using more of the skilled birth attendant services. As recommended by WHO, the majority of women had 4 or more ANC visits, and most made their first ANC visit in the first trimester. Most of the deliveries were done by unskilled personnel. Education of women, residence, parity, occupation, husband's education, and distance to health facility had a significant association with the use of skilled birth attendants. Finally, the findings show that urban residence and higher/vocational education are significantly associated with the uptake of the standard PNC package. To improve the utilization of maternal health care services, the Ministry of Health and Sanitation/Central Government should work with relevant stakeholders to formulate policies and design programs that target the advancement of education, rural infrastructure, and the empowerment of women. The ministry of health should Engage the research and scientific committee to undertake or facilitate further research programs to determine the healthcare factors that influence the uptake of MHC services. --- Availability of data and materials All data generated or analyzed during this study are included in this published article [and its supplementary information files]. --- Abbreviations --- Supplementary Information The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s12884-022-04597-z. --- Additional file 1. --- Authors' contributions Conceptualization and design: DMK, AAL, AM, EK, DAB. Data collection: DMK, JYK. Analysis and report writing: DMK, AAL, JYK. Drafting manuscript: DMK, DAB, AM, JYK, EK, AAL. All authors read and approved the final manuscript. • fast, convenient online submission • thorough peer review by experienced researchers in your field --- • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year --- • At BMC, research is always in progress. --- Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research? Choose BMC and benefit from:? Choose BMC and benefit from: --- Declarations Ethics approval and consent to participate Ethical approval was sort from the Sierra Leone Ethics and Scientific Committee. Permission was also sorted from the Kailahun Regional Health Directorate. Informed consent was sort from the respondents. For respondents below 18 years, the accent was granted by the parent or guardian. They were assured of confidentiality and informed of the purpose of the study. The information was stored without the names of the respondents in a folder that is only accessible to the research team. For participants who could not read, the consent was read and explained to them in the presence of an impartial witness or stakeholder in the community who also appended his signature on the consent form in accordance with the local IRB regulation. All study methods were carried out following relevant guidelines and regulations. --- Consent for publication Not applicable. --- Competing interests The authors declare that they have no competing interest. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: Ascertaining the key determinants of maternal healthcare service utilization and their relative importance is critical to priority setting in policy development. Sierra Leone has one of the world's highest maternal death ratios in the context of a weak health system. The objectives of this study were to determine; the level of utilization of Antenatal Care (ANC), Skilled Delivery Attendants (SDA), Postnatal Care (PNC) services, and factors that influence the utilization of these services.We conducted a community-based cross-sectional study involving 554 women of reproductive age (15-49 years) who had at least one delivery in the last 3 years and lived in the Kailahun District, Sierra Leone from November 2019 to October 2020. Data were analysed using analysed using bivariate, multivariate and multinomial logistic regression models.The median age of respondents was 25 years (Q1 = 17 years, Q3 = 30 years). Eighty-nine percent (89%) had 4 or more ANC visits. Only 35.9% of women were delivered by SDA. Women residing in urban areas had over six-fold increased odds of utilizing SDA as compared to women residing in rural areas (AOR = 6.20, 95% CI = 3.61-10.63). Women whose husbands had a primary level of education had 2.38 times increased odds of utilizing SDA than women whose husbands had no education (AOR = 2.38, 95% CI = 1. 30-4.35). Women that walked longer distances (30-60 min) to seek healthcare had 2.98 times increased odds of utilizing SBA than those that walked shorter distances (< 30 min) (AOR = 2.98, 95% CI = 1.67-5.33). Women who had a secondary/vocational level of education had 2.35 times increased odds of utilizing the standard PNC category as compared to those with no education (OR = 2.35, 95% CI = 1. 19-4.63).The majority of women had 4 or more ANC visits yet the use of skilled birth attendants was low. Urban residence and education were significantly associated with the use of the standard PNC category. To improve the utilization of maternal health care services, national healthcare policies should target the advancement of education, train skilled Maternal Healthcare (MHC) attendants, rural infrastructure, and the empowerment of women.

activity/ experience, was recently suggested as a stand-alone phenomenon differentiated from loneliness and negative solitude. As loneliness was previously found to have adverse implications for mental health, the present study examined whether the ability to engage in positive solitude can moderate the harmful effect of loneliness on depressive symptoms. The sample consisted of 520 community dwelling older adults in Israel aged [68][69][70][71][72][73][74][75][76][77][78][79][80][81][82][83][84][85][86][87]. Participants answered an online questionnaire through a survey company (Ipanel) assessing their background characteristics, depressive symptoms, loneliness, and positive solitude. The results show that loneliness was positively associated with depressive symptoms, whereas positive solitude was negatively associated with depressive symptoms. Furthermore, positive solitude moderated the relationship between loneliness and depressive symptoms, such that higher levels of positive solitude weakened this association. The findings indicate that positive solitude may serve as a buffering factor for mental health among older adults by augmenting coping with the adverse outcomes of loneliness. The results provide insight for tailoring future treatment interventions focusing on positive solitude to enhance mental health among older adults. Abstract citation ID: igad104.1059 --- SOLITUDE LINKS WITH INDIVIDUAL AND RELATIONAL WELL-BEING: EVIDENCE FROM DYADIC DAILY LIFE ASSESSMENTS Elizabeth Zambrano Garza 1, Rachel Murphy 1, Maureen Ashe 1, Kenneth Madden 1, Denis Gerstorf 2, Christiane Hoppmann 1, and Theresa Pauly 3, 1. University of British Columbia, Vancouver, British Columbia, Canada, 2. Humboldt University of Berlin,Berlin,Berlin,Germany,3. Simon Fraser University,Vancouver,British Columbia,Canada Solitude, defined as physical aloneness or the absence of social interactions, has been associated with both positive and negative outcomes. Older adults spend a significant amount of their waking time by themselves (exacerbated during the pandemic), but they may be better equipped to regulate the negative emotions that are often linked with being alone. Additionally, solitude may provide a balance between social connectedness and autonomous needs. Using pandemic, end-of-day daily diary data across 10 days from 136 older Canadian adults and a close other of their choice (59% spouses, M = 66.49 years, SD = 13.26 range: 18-87 years, 88% White, 62% women), this project aims to examine associations of actor and partner solitude with relationship quality and daily affect. Multilevel models revealed that participants who reported more voluntary solitude experienced more daily positive affect and those who reported more negative solitude experienced both more negative affect and less positive affect at both daily and overall levels. We also found daily partner effects such that more voluntary solitude of the partner was associated with less actor negative affect and negative solitude of the partner being associated with less actor positive affect and more negative affect over and above actor effects. Finally, when partners reported more voluntary solitude, participants reported higher relationship support. Findings shed light on the benefits and drawbacks of how solitude is experienced in older adulthood, highlighting the influence social partners affect quality and relationship satisfaction. --- Abstract citation ID: igad104.1060 --- ALONE WITH MY THOUGHTS: USING NATURAL LANGUAGE PROCESSING TO DISTINGUISH LONELY FROM CALM ALONENESS Jennifer Lay, University of Exeter, Exeter, England, United Kingdom Time spent alone is common in daily life, and particularly in old age; it may at times be experienced as lonely, and at times as calming (positive solitude). The use of machine learning to interpret human language (e.g., natural language processing or NLP) has been growing in psychology, and NLP has been used to distinguish experiences of loneliness from positive solitude in social media data (e.g., tweets). How effective might such NLP techniques be for classifying naturally-occurring aloneness experiences in older adults' everyday lives? The present study uses 1,546 thought samples collected from 133 adults aged 18-85 (M = 49.6 years; 73% female) at moments when they were alone over a 1-3 week period; participants reported their current thoughts (open-ended) and affective states (lonely, calm) at each assessment. Support vector machines were then used to classify moments of aloneness as being lonely (vs. not lonely) and calm (vs. not calm) based on the words participants used when describing their thoughts. When compared against participants' actual reported affective states, these classifications achieved up to 67% accuracy for "lonely" and 79% accuracy for "calm". "Lonely" thought reports included the word "work" more frequently, whereas "calm" thought reports more frequently included the words "reading" and "think/ thought", suggesting individuals engage in quiet or contemplative activities during positive solitude. Age differences therein are also examined. Findings suggest that NLP may be a useful tool for identifying older adults at risk of loneliness (and its negative health implications) by using brief reports of their everyday thoughts. --- SESSION 3280 (SYMPOSIUM) Abstract citation ID: igad104.1061 --- ACTIVE ENGAGEMENT AFTER RETIREMENT Chair: Dannii Yeung Co-Chair: Edwin Chung Retirement has been shown to accelerate the declines in physical and mental health (Dave et al., 2008) and cognitive functioning (Celidoni et al., 2017;Xue et al., 2018) due to the identity disruption and removal of a cognitively simulating environment after leaving the workforce. This necessitates a need to systematically uncover factors that could potentially buffer the negative impacts brought by retirement. This symposium aims to identify the types of self-initiated activities that can maintain post-retirement well-being through research conducted in Hong Kong and Japan. Specifically, the first paper from Chung reveals that retirees engaging in postretirement work exhibit lower life satisfaction but higher cognitive functioning than those without work. Similarly, the second paper from Katagiri also shows that work has negative impacts on life satisfaction, but such effect is buffered by civic participation. The third paper from Ho suggests that the beneficial effects of ICT use on personal mastery is more prevalent in retirees than in non-retirees. The fourth paper from Lin discloses that participation in cognitively demanding volunteering activities is associated with better cognitive and psychological functioning. These four presentations unveil various types of post-retirement activities that are beneficial to retirees, including re-engagement in the workforce, the use of technology for leisure and financial purposes, and participation in cognitively stimulating volunteering. Given that most of the pre-retirees do not have much planning on social life arrangement, findings of this symposium provide practical implications to improve psychological adjustment to this critical life event. Abstract citation ID: igad104.1062 --- THE EFFECTS OF POSTRETIREMENT WORK ON WELL-BEING OF RETIRED ADULTS Edwin Chung, Gloria Lin, and Dannii Yeung, City University of Hong Kong, Hong Kong, Hong Kong There has been evidence that a growing number of retirees are reentering the labor force (DeSilver, 2016;Greenwald et al., 2017), possibly due to financial needs and seeking continuity in their preretirement job (Burkert & Hockfellner, 2017). However, the effects of post-retirement work on retirees' well-being remain under-examined. This cross-sectional study thereby aims to examine the effects of post-retirement work in a sample of 768 younger and 530 older Hong Kong Chinese retirees (Mage = 65.1, SD = 2.72, range = 60 -69, and Mage = 75.2, SD = 4.50, range = 70 -95, respectively). Participation in post-retirement work, life satisfaction, cognitive functioning and hand-grip strength were measured. The results of two-way between-subject robust ANOVAs demonstrated that retirees engaging in postretirement work exhibited lower life satisfaction (Q = 5.82, p =.02) but higher cognitive functioning (Q = 5.29, p =.02) than those without post-retirement work. Furthermore, as compared with younger retirees, older retirees engaging in post-retirement work exhibited higher cognitive functioning (Q = 3.66, p =.05) and stronger hand-grip strength (Q = 4.21, p =.04). These results remain significant even after controlling for sex, education, socioeconomic status, perceived health, and financial control. The findings of this study thus reveal the beneficial effects of post-retirement work on cognitive and physical health, but not on life satisfaction. The negative effect of post-retirement work on subjective well-being will be discussed in relation to the unsatisfactory working environment or employee benefits received by older workers in Hong Kong.

The present study describes the development of a new questionnaire for measuring positive solitude. First we will discuss the conceptualization of positive solitude, and refer to the contradictions that exist in the literature regarding this concept. We will then present the theoretical background on which we based our questionnaire, and some empirical findings regarding its utility. We will describe the 9-item tool and its relationships to positive and negative measures of well-being and mental health. Furthermore, we will discuss findings from a Holocaust survivors study showing the role of positive solitude on posttraumatic symptoms. We will conclude this presentation by discussing several future directives, some of which that will be discussed in other presentations of this symposium.

Background: Little is known about the link between perceived neighborhood walkability and prevalence of chronic disease. Even less is known regarding this association among Hispanic/Latino adults, despite exhibiting high rates of chronic diseases. Stress due to racial discrimination is a harmful social determinant of health in Hispanics/Latinos. Having both low perceived neighborhood walkability and high racial discrimination stress may exacerbate the chronic disease status of Hispanics/Latinos. Among a U.S. national sample of Hispanic/Latino adults, this cross-sectional study aims to examine (1) the associations among overall perceived neighborhood walkability, racial discrimination stress, and having a chronic health condition; and (2) whether overall perceived neighborhood walkability moderates the hypothesized association between racial discrimination stress and having a chronic health condition. Methods: In January 2018, 798 Hispanic/Latino adults (M age = 39.7 years, SD = 15.1; 58.6% female; 70.0% U.S. born; 52.0% Mexican/Mexican American) responded to a survey via Qualtrics Panels. Surveys included the Neighborhood Environment Walkability Scale-Abbreviated, Hispanic Stress Inventory-2, and self-reported presence/absence of chronic health conditions (e.g., hypertension, heart disease). A logistic regression was conducted testing for the moderation of the main effect of racial discrimination stress on the presence of a chronic health condition by overall perceived neighborhood walkability. Results: After controlling for age, body mass index, and income, racial discrimination stress was inversely associated with overall perceived neighborhood walkability (b = -0.18, p <unk> 0.001) and positively associated with having a chronic health condition (OR = 1.02; 95% CI [1.00, 1.03]). While overall perceived neighborhood walkability was not associated with having a chronic health condition, perceived crime safety was inversely associated with having a chronic health condition (OR = 0.94; 95% CI [0.89, 0.99]). Perceived crime safety moderated the positive association between discrimination stress and having a chronic health condition, such that the association was only significant among those who perceived their neighborhood to be less safe (<unk> = -0.004, 95% CI [-0.01, -0.00]). --- INTRODUCTION In the United States, chronic diseases including cancer, coronary heart disease, and type 2 diabetes are the leading causes of death among Hispanics/Latinos (1,2). Type 2 diabetes is also more prevalent among Hispanics/Latinos (17%) compared with non-Hispanic Whites (8%) (3). Additionally, Hispanics/Latinos are more likely to develop type 2 diabetes at an earlier age and have more severe diabetes-related complications than the U.S. adult population overall (3). Neighborhood walkability, which is how conducive the area around one's home is to walking, has been linked to prevalence of chronic disease (4)(5)(6)(7)(8). Most work on neighborhood walkability has employed objective measures such as land use, population density, crime rates, and/or proximity to parks (9). A 10-year longitudinal study employing objective measures found that, as neighborhoods became more walkable over time, community-wide diabetes prevalence and cardiovascular disease risk (e.g., high blood pressure and cholesterol) declined (8). Others have reported similar associations using objective measures of neighborhood walkability (4)(5)(6). The facilitation of physical activity is one behavioral mechanism identified as connecting greater objectively-measured walkability with lower prevalence of chronic disease (4,(9)(10)(11)(12). In comparison, perceived neighborhood walkability is defined as the summation of a person's subjective evaluations of the environment around their home such as their perceptions of aesthetics, built environment features (e.g., presence of crosswalks), and/or safety from crime and traffic (9). A large, multi-country study found certain domains within perceived neighborhood walkability were associated with body mass index (BMI) (7). Most notably, greatest perceived safety from traffic and crime, and nearness to destinations (e.g., stores) were strongly associated with lower BMI (7). The association between perceived neighborhood walkability and chronic diseases is unknown. However, some evidence has suggested that one's perception of neighborhood walkability may have more influence on one's physical activity engagement than objective measures of neighborhood walkability (13,14). Furthermore, of the extant literature on the topic, Hispanic/Latino adults make up a small percentage of the overall participants investigated, despite having some of the highest rates of chronic diseases (6,8,11). When Hispanics/Latinos are included in studies of neighborhood walkability, they are often compared to their non-Hispanic White peers, hindering our ability to capture within ethnic group variability in experiences, behaviors and health outcomes (15). These data are necessary to inform culturally tailored, community-based interventions to better address disparities in chronic diseases among Hispanic/Latino adults in the United States (16,17). A particularly salient social determinant of health among people of color in the United States is racial discrimination (18)(19)(20). The Psychosocial Stress Model explains that the stress associated with institutional racism and individual-level experiences with racism lead to health disparities, such as those documented among Hispanic/Latino adults (3,18,21). Institutional racial discrimination refers to systems or structural differences that manifest in inequitable distributions of resources and opportunities for people of color. Individuallevel experiences of racial discrimination are defined as one's experiences of being treated more poorly than others because of one's race or ethnicity. Empirical evidence suggests that heightened stress due to experiences of racial discrimination is linked to higher rates of chronic diseases by way of several pathways, such as heightened cortisol production, accelerated cellular aging, and poorer health behaviors, including physical activity avoidance, smoking, and greater alcohol consumption (20,(22)(23)(24). Meta-analysis results suggest that the inverse association between experiences of racial discrimination and general health may be particularly robust for Hispanic/Latino adults (25). Since experiences of discrimination often take place in public settings (26), including one's neighborhood, it is plausible that perceived neighborhood walkability and racial discrimination stress interact to increase risk for poor health behaviors and associated chronic disease risk. Additionally, neighborhood walkability can be described as an example of institutional racism, such that Hispanics/Latinos disproportionately live in low socioeconomic areas, which is associated with lower objective and perceived neighborhood walkability (27,28). Theoretical models and empirical literature tend to suggest one's environment is a moderator in the association between psychosocial factors and health/health behaviors (18,29,30). However, the association between perceived neighborhood walkability and racial discrimination stress and how they interact in their associations with chronic diseases among Hispanics/Latinos in the United States are unknown. Among a national sample of U.S. Hispanic/Latino adults, this study aims to (1) test the associations among overall perceived neighborhood walkability, racial discrimination stress, and having a chronic health condition; and (2) examine whether overall perceived neighborhood walkability moderates the association between racial discrimination stress and having a chronic health condition. Informed by the aforementioned literature, hypotheses to the study aims are outlined as follows. Regarding aim 1, it is hypothesized that there will be a significant inverse association between racial discrimination stress and overall perceived neighborhood walkability (26,28,31); a significant positive association between racial discrimination stress and having a chronic health condition (18,20,(22)(23)(24)(25); and a significant inverse association between overall perceived neighborhood walkability and having a chronic health condition (7,11). Regarding aim 2, it is hypothesized that overall perceived neighborhood walkability will significantly moderate the association between racial discrimination stress and having a chronic health condition (18,29,30), such that higher overall perceived neighborhood walkability will reduce the strength of the association between racial discrimination stress and having a chronic health condition, adjusting for age, BMI, annual household income, acculturation, U.S. region of residence, and health insurance status (32)(33)(34)(35)(36)(37)(38)(39)(40)(41). The findings of this study advance the literature by elucidating the associations among social and environmental factors and the presence of a chronic health condition among Hispanic/Latino adults, a sizable sub-set of the U.S. population at heightened risk for chronic diseases (3,21). A better understanding of these associations could inform culturally tailored, communitybased strategies to better promote healthy behaviors and prevent chronic diseases among Hispanics/Latinos in the United States (16,17,42). --- MATERIALS AND METHODS --- Data Collection In January 2018, Qualtrics Panels was used to recruit Hispanic/Latino adults to complete a 197-question electronic survey in English or Spanish, which assessed their demographic information, health behaviors, physical and mental health conditions, and perceptions of their social and physical environments. Measures in the survey that were not already validated in Spanish were translated forward and backward to Spanish from English by two bi-lingual research assistants. Eligible respondents were 18 years or older, identified as Hispanic or Latino, resided in the United States, and were fluent in English or Spanish. Qualtrics Panels partners with over 20 online panel providers, namely market research panels, to supply a nationwide sample of potential participants who fit the eligibility requirements of this cross-sectional study (43). Potential research participants were contacted by Qualtrics Panels via an email invitation. To avoid self-selection bias, the specific details of the study were not included in the email invitation. Those who expressed interest in participating were sent study details and underwent informed consent protocols in line with the committee responsible for human experimentation (institutional and national) and with the World Medical Association's Declaration of Helsinki. Specifically, prior to taking the survey, all participants provided their consent by selecting a radial button that read, "I have read the contents of this consent form. I give my consent to participate in this study, " which appeared at the bottom of the electronic informed consent form. Written documentation of consent was waived. Participants who completed the survey, received a pre-determined amount of points (stated in the email invitation to participate in the study). The amount of points was determined by the length of the survey, participant's specific panelist profile, and target acquisition difficulty. The points earned for completion of the survey can be redeemed for cash, airline miles, gift cards, points on online games, sweepstakes entrances and vouchers. This research protocol was approved by the Institutional Review Board at the University of Oregon. The dataset analyzed for this study can be found in the Harvard Dataverse, (44). --- Measures Demographic characteristics of the respondents were assessed using the Demographic and Health Information Data Questionnaire (45,46). Demographic characteristics reported in this study include respondents' age, height, weight, gender, ethnic origin, whether English was their first language, nativity, acculturation, highest level of education, employment status, U.S. region of residence, relationship status, health insurance status, and annual household income. Response options for all demographic characteristics appear in Table 1. BMI was calculated using the following equation: 703 <unk> weight (lbs) /[height (in)] <unk> 2 (47). Seven respondents were excluded from the analyses because of seemingly implausible BMI values (<unk>12 or >65). Annual household income was assessed by asking respondents to select the total combined income that is made yearly by all working members of their household from 11 U.S. dollar ranges. Response options were collapsed into: <unk>$10,000-$29,000, $30,000-$69,999, and $70,000-$100,000+. Acculturation was assessed with The Abbreviated Multidimensional Acculturation Scale, which includes 6 U.S. cultural identity items (e.g., "I think of myself as being U.S. American"), nine English language competence items (e.g., "How well do you speak English on the phone?"), and six U.S. cultural competence items (e.g., "How well do you know popular American television shows?") (48). U.S. cultural identity items were assessed on a Four-point-Likert scale with response options ranging from 1 ("strongly disagree") to 4 ("strongly agree"). English language competence and U.S. cultural competence were assessed on a Four-point-Likert scale with response options ranging from 1 ("not at all") to 4 ("extremely well"). The Abbreviated Multidimensional Acculturation Scale has been shown to produce reliable and valid responses among Hispanics/Latinos (48). A mean was calculated to create a total U.S. acculturation score from all of the items within the U.S. cultural identity, English language competence, and U.S. cultural competence sub-scales. A higher mean score denotes a greater degree of acculturation to the United States. The independent variable, racial discrimination stress, was measured by the total score of the Discrimination Stress subscale (11 items; e.g., "I was discriminated against because of my customs and cultural celebrations") from the Hispanic Stress Inventory-2. The Discrimination Stress sub-scale assesses psychosocial stress from discriminatory experiences related to being Hispanic. For each item, respondents were asked if they have experienced the reported stressor (Yes/No), and if they had experienced the stressor, they were asked to rate how stressful the event was on a five-point-Likert scale ranging from 1 ("not at all worried/tense") to 5 ("extremely worried/tense"). Possible scores ranged from zero to 55 and higher scores indicate greater racial discrimination stress. The Hispanic Stress Inventory-2 sub-scales have been shown to produce reliable and valid responses in Spanish and English among a diverse sample of Hispanic/Latino adults in the United States (49,50). In this sample, the Discrimination Stress sub-scale had good reliability (<unk> =.933). The dependent variable, presence of a chronic health condition, was assessed by one question from the Demographic and Health Information Data Questionnaire (45,46). Respondents were asked to indicate all of the health conditions they currently have from the following: high cholesterol, overweight/obesity, heart disease, cancer (specify the type), high blood pressure, type 2 diabetes, other (specify), or to select "I do not have any of these health conditions." The hypothesized moderator, overall perceived neighborhood walkability, was measured using a total sum score of four subscale scores from the Neighborhood Environment Walkability Scale-Abbreviated (NEWS-A) including: infrastructure for walking/cycling (six items; e.g., "There are sidewalks on most of the streets in my neighborhood"), aesthetics (four items; e.g., "There are trees along the streets in my neighborhood"), traffic safety (three items; e.g., "The speed of traffic on most nearby streets is usually slow (30 mph or less)"), and crime safety (three items; e.g., "The crime rate in my neighborhood makes it unsafe to go on walks during the day") (51). Respondents chose from four-point Likert scale response options ranging from 1 ("strongly disagree") to 4 ("strongly agree") for each item. Five items were reverse coded to align with the direction of the other items. Possible scores ranged from 16 to 64 and higher scores indicate greater overall perceived neighborhood walkability. The scale has demonstrated validity and reliability in English and Spanish (52,53). In this sample, the overall NEWS-A had adequate reliability (<unk> = 0.761). --- Analyses Descriptive statistics were conducted to review missing cases and response distributions for scale variables with a skew or kurtosis outside of <unk>2 (54). Preliminary analyses also included a Pearson correlation in order to examine potential multicollinearity (r > <unk> 0.80) between key study variables and independent-samples t-tests to assess unadjusted associations between key study variables and the dependent variable (55). Regarding the dependent variable, having "overweight/obesity" was removed as one of the chronic conditions. Instead, BMI was included as a covariate in the regression models because of its association with chronic disease risk (39). Respondents endorsed having up to four of the six included chronic health conditions. Of those who endorsed having at least one of the health conditions, 60.22% reported having only one of the conditions followed by 23.48% with two. Given this distribution, responses were dichotomized into whether the respondent endorsed currently having either any or none of the health conditions. Prior research also supports treating presence of a chronic health condition as a binary term (36,56), given meaningful differences between the two groups in health service use (e.g., emergency room visits) and general medical expenditures (36). To address study aim 1, one multiple linear regression model and two binary logistic regression models were conducted. Age, annual household income, BMI, acculturation, U.S. region of residence, and health insurance status were initially included in these models as covariates because of their associations with having a chronic health condition (32)(33)(34)(35)(36)(37)(38)(39)(40)(41). In pursuit of parsimonious models, covariates were removed that had non-significant (p <unk> 0.05) regression coefficients in all of the regression models conducted for study aims 1 and 2, and also had no notable effects on the models' variance explained or results when they were removed (57). First, the association between overall perceived neighborhood walkability and racial discrimination stress, adjusting for the covariates, was assessed in a multiple linear regression model. Second, the main effects of racial discrimination stress on presence of a chronic health condition, adjusting for the covariates, was investigated using binary logistic regression. Third, the main effects of overall perceived neighborhood walkability on the presence of a chronic health condition, adjusting for the covariates, was conducted using binary logistic regression. The binomial assumption was assumed because there was no reason to believe there was dependence among observations or that the sample was nonrandom (58). For study aim 2, overall perceived neighborhood walkability and racial discrimination stress were grand-mean centered and an interaction term was created. A binary logistic model was conducted that included: the covariates, centered overall perceived neighborhood walkability, centered racial discrimination stress, and the interaction term. All statistical analyses were conducted in IBM SPSS Statistics for Windows, Version 26 (59). The moderation plot was conducted in R, Version 3.6.1 using the interactions package (60,61). --- Exploratory Analyses For post hoc exploratory purposes, aims 1 and 2 were reexamined using the individual sub-scales of the NEWS-A. Infrastructure for walking/cycling (<unk> = 0.764), aesthetics (<unk> = 0.799), and crime safety (<unk> = 0.912) demonstrated adequate to good reliability in this sample. Traffic safety demonstrated poor reliability (<unk> = 0.258) in this sample and was thus not included in the exploratory analyses. Preliminarily, Pearson correlations were conducted to assess associations among the three sub-scales. The same statistical procedures were conducted as described earlier for the study aims, except they were repeated three times, once for each NEWS-A sub-scale. --- RESULTS --- Sample The survey respondents included 798 Hispanic/Latino adults 18-81 years old (M = 39.7 <unk> 15.1 years; BMI M = 28.30 <unk> 7.3 kg/m 2 ; Table 1). Approximately 20% of respondents completed the survey in Spanish, 80% in English. The majority of the sample identified as women (58.6%); Mexican or Mexican American (52%); native English speakers (65.4%); U.S. born (70.0%); educated through 2 years of college or beyond (62%); and employed at least part time (64%). Respondents resided in all major regions of the United States, with the largest portion living in the South (39.6%). Additionally, the largest proportions of respondents had a least one chronic health condition (45.4%); were married or in a civil union (45.6%); had public or private health insurance (87.2%); and an annual household income of $30,000-69,999 (37.6%). --- Preliminary Findings In unadjusted associations, there was a significant inverse association between overall perceived neighborhood walkability and racial discrimination stress (r = -0.166, p <unk> 0.01). The mean scores of overall perceived neighborhood walkability (t (787) = -0.695, p = 0.487) and racial discrimination stress (t (792) = -0.963, p = 0.336) were not significantly different for those with or without a chronic health condition. --- Study Aim 1 Results The covariates retained in aim 1 and aim 2 regression models were age, BMI, and annual household income. There was a significant and inverse association between overall perceived neighborhood walkability and racial discrimination stress (b = -0.18, p <unk> 0.001; R 2 = 0.07). Table 2 provides an overview of results from the logistic regression analyses. In Model 1, there was a significant and positive association between racial discrimination stress and the presence of a chronic health condition (Odds Ratio [OR] = 1.017; 95% Confidence Interval [CI] [1.004, 1.030]). In Model 2, there was no significant association between overall perceived neighborhood walkability and the presence of a chronic health condition (OR = 0.993; 95% CI [0.72, 1.015]). --- Study Aim 2 Results The full model results displayed in Model 3 within Table 2 show that the interaction between racial discrimination stress and perceived walkability on having a chronic health condition was not significant (OR = 0.999; 95% CI [0.997, 1.001]), suggesting the association between racial discrimination stress and health did not vary by overall perceived neighborhood walkability. Model 3 explained a total 17.8% of the variance in the presence of a chronic health condition. --- Exploratory Results Unadjusted bivariate associations among the three NEWS-A sub-scales showed crime safety was inversely associated with infrastructure for walking/cycling (r = -0.106, p <unk> 0.01) and not significantly associated with aesthetics (r = 0.059, p = 0.09). Infrastructure for walking/cycling was positively associated with aesthetics (r = 0.520, p <unk> 0.01). Table 3 shows the association between each NEWS-A subscale and racial discrimination stress, adjusted for age, BMI, and annual household income. Only crime safety was significantly and inversely associated with racial discrimination stress (b = -0.27, p <unk> 0.001). Among the three NEWS-A sub-scales, there was only a significant and inverse association between crime safety and the presence of a chronic health condition (OR = 0.936; 95% CI [0.887, 0.988]; Table 4). There was also a significant moderating effect of crime safety on the association between racial discrimination stress and presence of a chronic health condition. Specifically, there was a significant and positive association between discrimination stress and having a chronic health condition, only among those who perceived their neighborhood to be less safe (<unk> = -0.0043, SE = 0.0021, CI[-0.0084, -0.0002], p <unk> 0.05; Figure 1). --- DISCUSSION The present study examined (1) the associations among overall perceived neighborhood walkability, racial discrimination stress, and having a chronic health condition; and (2) whether overall perceived neighborhood walkability moderated the hypothesized association between racial discrimination stress and having a chronic health condition among a U.S. national sample of Hispanic/Latino adults. Study findings advance the understanding of how experiences such as, how individuals perceive the safety of their neighborhoods and the degree to which individuals experience racial discrimination stress, vary among Hispanics/Latinos in the United States and how these experiences are associated with having a chronic health condition. The study findings can inform next steps in research and practice that aim to eliminate current chronic disease disparities among Hispanic/Latino adults and details for which are provided below (3,21). As hypothesized, those who reported experiencing more racial discrimination stress had higher odds of having a chronic health condition after adjusting for age, BMI, and annual household income. This finding is consistent with The Psychological Stress Model (18) and extant literature among Hispanics/Latinos and other racial/ethnic miniories (19,23,25,49). While the increased odds for having a chronic health condition among those who report more racial discrimination stress was marginal in this study, alongside the extant literature, the finding reinforces the importance of reducing racial discrimination against Hispanics/Latinos in the United States in order to promote health equity. Furthermore, as hypothesized (26,28,31), an inverse association, although weak, was found between overall perceived neighborhood walkability and racial discrimination stress. Overall perceived neighborhood walkability is a total of four individual sub-scales. Exploratory results indicated perceived safety from crime, not perceived infrastructure for walking/cycling and aesthetics of one's neighborhood, was the driver of the inverse association found between overall perceived neighborhood walkability and racial discrimination stress. These findings not only advance our understanding of the association between these environmental and social factors among U.S. Hispanic/Latino adults, they are also novel to the broader literature not focused on Hispanics/Latinos. In related literature, a systematic review found consistent support for the inverse association between objectively measured neighborhood walkability and general psychosocial distress (62). In addition to focusing on objective vs. perceived neighborhood walkability, the reviewed studies did not assess racial discrimination stress specifically and only one of the studies, by Brown and colleagues (63), was strictly among U.S. Hispanic/Latino adults (62,63). Brown et al.'s (63) study involved low socioeconomic, U.S. Hispanic/Latino, older adults and found that front porches were positively associated with social support, which was associated with lower psychological distress (63). The presence of front porches is most similar to items in the infrastructure for walking/cycling sub-scale within the NEWS-A. This association between infrastructure for walking/cycling and racial discrimination stress was not identified in the present study, which could be due to the use of different measures of perceived neighborhood walkability between the studies and substantial demographic differences between the study samples. Future research could investigate social support and other potential links connecting perceived neighborhood crime safety and racial discrimination stress among Hispanic/Latino adults. Contrary to the hypothesis, overall perceived neighborhood walkability was not associated with having a chronic health condition. However, exploratory findings showed that perceived neighborhood safety from crime was inversely associated with having a chronic health condition. Both of these findings contribute toward filling the gap in knowledge regarding perceived (vs. objective) neighborhood walkability and chronic diseases (4-6, 8). There is a precedent for perceived neighborhood safety playing a pivotal role in behavioral and health risk factors for chronic diseases in the literature. For example, another study of Hispanics/Latinos found infrastructure for walking/cycling was positively associated with physical activity, but only when perceived crime was low (64). Also, a multi-country study by De Bourdeaudhuij and colleagues (7) found perceived neighborhood safety from traffic and crime, and nearness to destinations were inversely associated with BMI, adjusting for relevant demographic factors (7). Where BMI was the dependent variable in De Bourdeaudhuij's (7) study, BMI was a covariate in the present study. The latter approach is advantageous when aiming to reduce chronic disease rates among Hispanics/Latinos, as higher BMI is just one of many risk factors for chronic diseases (65). Overall perceived neighborhood walkability did not moderate the association between racial discrimination stress and presence of a chronic health condition, contrary to the hypothesis. However, exploratory findings did identify a significant moderating effect of perceived crime safety on the association between racial discrimination stress and presence of a chronic health condition. Specifically, the positive association between racial discrimination stress and presence of a chronic health condition is significant only among those who report lower perceived neighborhood safety from crime. This finding is consistent with theoretical models and empirical literature that have characterized the environment as a moderator of the association between psychosocial factors and health/health behaviors (18,29,30). The findings also highlight the importance of studying and targeting the perceived walkability of neighborhoods in more specific rather than general ways. Examining the study aims by each NEWS-A sub-scale brought clarity where overall perceived neighborhood walkability could not. This study's findings suggest perceived neighborhood safety matters, not infrastructure and aesthetics, when it comes to racial discrimination stress and having a chronic health condition among Hispanics/Latinos. Low perceived neighborhood safety from crime could reflect institutional racism, as Hispanics/Latinos are more likely than non-Hispanic Whites to be victims of violent crimes and more likely to live in poverty, which is associated with higher crime rates (66,67). Improving safety in communities where Hispanic/Latino residents report greater safety concerns may have beneficial effects on racial discrimination stress and health among Hispanic/Latino adults. Testing such an intervention in a longitudinal study is a recommended next step in research. This study responds to calls in the literature to investigate innovative explanations for health disparities by testing interactions of the physical and social environments (68,69). A study that examined the moderating role of the environment on racial discrimination stress and health among African Americans found that discrimination experienced in the previous year was positively associated with cortisol concentration among those residing in neighborhoods with more White residents (69). Building on the present study's findings by including an additional aspect of the social environment, such as social capital (70), perceived neighborhood social cohesion (71), or neighborhood racial makeup (69), may explain more of the total variance in presence of a chronic health condition among Hispanic/Latino adults. The study sample reflected the Hispanic/Latino population in the United States well on a number of sociodemographic characteristics including proportions by ethnic origins, nativity, employment status, relationship status, and health insurance status (72,73). The sample appears to over represent women (sample 58% vs. U.S. 49%) and those who have completed 2 years or more of college (sample 62% vs. U.S. 41%), and underrepresent those living in the Western region of the United States (sample 23% vs. 41%) (73,74). These disproportionate representations in the sample could have implications for the findings, but explanations follow that propose the implications may be minimal. Suspecting the study results could vary by gender, post hoc analyses were conducted and the study findings did not change when examined by women and men. Additionally, those with more education are more likely to have a primary healthcare provider and more likely to effectively manage chronic health conditions, like hypertension, than those less educated (75,76). Thus, this study's findings could reflect the experiences of a healthier group of Hispanics/Latinos than the average national cross-section of Hispanics/Latinos. A RAND Corporation report on nationally representative data from the Medical Expenditure Panel Survey from 2014 shows that 49% of Hispanic/Latino adults in the United States have at least one chronic health condition. This is 3.6% more than respondents endorsed in this study, and the two studies generally assessed the same conditions. Finally, the Western region of the United States, California in particular, is home to the largest proportion of the U.S. Hispanic/Latino population. One national survey found that the West generally fares better on indices of residential, racial integration than other U.S. regions, but the state of race relations between non-Hispanic Whites and Hispanic/Latinos were similar to the rest of the country (77). Region was also included in the initial analyses as a covariate and removed only after it was determined to not be an explanatory contributor to the models. --- Strengths and Limitations The most notable strengths of this study are the several novel findings that advance the literature and inform next steps in practice and research that aim to address health disparities among Hispanic/Latino adults in the United States. Additional study strengths are that the survey was provided in English and Spanish and the sample reflected the U.S. Hispanic/Latino population well on a number of characteristics; both increase generalizability of the study results. A large sample size of Hispanic/Latino adults provides ample power to test for moderation, which is often underpowered in the literature (78), and allows for the examination of within ethnic group variability in experiences and chronic health conditions. Lastly, there are also study limitations. The cross-sectional design limits the study interpretations to associations rather than causal inferences. The inability to measure survey response rate using Qualtrics Panels and the un-measured differences between those who are more and less likely to participate in electronically delivered surveys makes the study vulnerable to selection bias. Although, a meta-analysis by Walter and colleagues (79) did find similar internal reliability and external validity between online panel and other sampling techniques (79). The self-report measure of current chronic health conditions, as opposed to medically confirmed, likely underestimates the actual proportion of those who have chronic health conditions, as undiagnosed type 2 diabetes is common, especially among people of color in the United States (80). Future longitudinal studies should focus on disentangling mechanisms by chronic condition, as some mechanisms are shared and others are not. Finally, the racial discrimination stress measure was not specific to racial discrimination experienced in one's neighborhood, however, it does capture stress due to racial discrimination in the contexts of one's daily life, which would include the neighborhood in which one lives. --- CONCLUSION Novel insights of this study include the associations among perceived neighborhood walkability, racial discrimination stress, and having a chronic health condition among U.S. Hispanic/Latino adults, as well as evidence that perceived neighborhood safety plays a moderating role in the positive association between racial discrimination stress and having a chronic health condition. Improving safety in communities where Hispanic/Latino residents report greater safety concerns may have beneficial effects on racial discrimination stress and health among Hispanic/Latino adults. --- DATA AVAILABILITY STATEMENT The dataset analyzed for this study can be found in the Harvard Dataverse, https://doi.org/10.7910/DVN/NABLZX. --- ETHICS STATEMENT The study involving human participants was reviewed and approved by Institutional Review Board at the University of Oregon. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. --- AUTHOR CONTRIBUTIONS EB contributed to the conception

Hispanic/Latino Walkability, Stress, and Health Conclusions: Overall perceived neighborhood walkability was inversely associated with racial discrimination stress, but not associated with having a chronic health condition. Perceived neighborhood crime safety, but not infrastructure or aesthetics, matters when it comes to the link between racial discrimination stress and having a chronic health condition among Hispanics/Latinos.

daily life, which would include the neighborhood in which one lives. --- CONCLUSION Novel insights of this study include the associations among perceived neighborhood walkability, racial discrimination stress, and having a chronic health condition among U.S. Hispanic/Latino adults, as well as evidence that perceived neighborhood safety plays a moderating role in the positive association between racial discrimination stress and having a chronic health condition. Improving safety in communities where Hispanic/Latino residents report greater safety concerns may have beneficial effects on racial discrimination stress and health among Hispanic/Latino adults. --- DATA AVAILABILITY STATEMENT The dataset analyzed for this study can be found in the Harvard Dataverse, https://doi.org/10.7910/DVN/NABLZX. --- ETHICS STATEMENT The study involving human participants was reviewed and approved by Institutional Review Board at the University of Oregon. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. --- AUTHOR CONTRIBUTIONS EB contributed to the conception and design of the study, analysis and interpretation of data, and drafting of the full manuscript. NG and NK contributed to the conception and design of the study, analysis and interpretation of data, and manuscript revisions. All authors read and approved the submitted version. --- Conflict of Interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Hispanic/Latino Walkability, Stress, and Health Conclusions: Overall perceived neighborhood walkability was inversely associated with racial discrimination stress, but not associated with having a chronic health condition. Perceived neighborhood crime safety, but not infrastructure or aesthetics, matters when it comes to the link between racial discrimination stress and having a chronic health condition among Hispanics/Latinos.

Introduction In 2008, national crime arrest statistics indicated that youth under age 18 in the United States were responsible for 16.2% of all violent crime arrests defined as murder, forcible rape, robbery, and aggravated assault (Federal Bureau of Investigation, 2008). Self-reports of youth violence reveal many more offenses compared to official crime statistics. For instance, according to the Centers for Disease Control and Prevention (2009), 36% of high school students reported being in a physical fight one or more times in the year preceding the survey, and 18% reported carrying a weapon on one or more of the 30 days preceding the survey. Racial disparities in violence are well documented. According to the 2008 FBI crime data, Black juveniles are 5 times more likely to be arrested for violent crime than are White juveniles (Puzzanchera, 2009). However, such data relies solely upon arrest records, --- Income and Violence Much of the existing research on racial disparities in violence has pointed to economic variables that help partially explain Black-White differences in adolescent violence perpetration. Research suggests that higher income is associated with fewer violent behaviors among both Black and White adolescents (Blum et al., 2000;Hsieh & Pugh, 1993). However, Black adolescents are more likely to come from families with fewer economic resources and live in communities characterized by higher levels of disadvantage than their White counterparts (Jargowsky, 1996;Massey, 1990). Many other studies have also found that different levels across race of other measures of disadvantage (community disadvantage and cumulative disadvantage) can contribute to an explanation of the variation in racial disparity in violence (Choi, Harachi, Gillmore, & Catalano, 2006;Haynie, Silver, & Teasdale, 2006;McNulty & Bellair, 2003). --- Families and Violence Racial disparities in violent behavior may be the effect of Black families having fewer economic resources. Parents who lack adequate financial resources often find themselves with less time to engage in closely monitoring their children (Conger et al., 1992), and may use more power-assertive forms of discipline (Heimer, 1997), both of which have been shown to increase the probability that a child will associate with deviant peers and become involved in violent or delinquent behaviors. Researchers argue that while economic factors can help explain some of the differences between Blacks and Whites in violent offending, they often do not explain all of the disparity (Blum et al., 2000;Haynie et al., 2006). Researchers have identified many ways in which parenting practices impact problem behavior such as delinquency and violence (Hawkins et al., 1998;Henry, Tolan, & Gorman-Smith, 2001;Herrenkohl et al., 2000;Patterson & Stouthamer-Loeber, 1984;Windle et al., 2010). Many researchers have argued that negative parenting practices, including harsh or inconsistent discipline, directly influence the development of violent or delinquent behaviors (Hawkins et al., 1998;Henry et al., 2001;Herrenkohl et al., 2000;Patterson & Stouthamer-Loeber, 1984). Both overly harsh and lax discipline have been linked to a range of adolescent problem behaviors, including violence (Farrington, 1989;Hawkins et al., 1992). Researchers have often found that adolescents who have been exposed to harsh and violent forms of discipline are more likely to become violent as adolescents and young adults (Heimer, 1997). Additionally, parental history of violence has been associated with the violent behavior of children (Hawkins et al., 1992;Herrenkohl et al., 2000). Caregivers can also have a protective influence on their children. Research has found that positive parenting practices that include establishing clear guidelines for behavior, positive monitoring, and the delivery of fair and consistent consequences can act as a buffer, shielding an adolescent from involvement in violent or delinquent behavior (Gorman-Smith, Henry, & Tolan, 2004;Griffin, Botvin, Scheier, Diaz, & Miller, 2000;Henry et al., 2001;Herrenkohl et al., 2003). Recent research finds that families play an important role in establishing guidelines for behaviors, in monitoring behaviors, and providing both positive and negative consequences for behaviors across races (Windle et al., 2010). Bonding or attachment to a primary caregiver is also an important family protective factor (Catalano & Hawkins, 1996). This argument is consistent with Hirschi's (1969) control theory-children do not want to damage the bonds with parents and therefore avoid situations which would vary from parental expectation for child behavior. Mack and colleagues (2007), in support of Hirschi's control theory, found that maternal attachment was a stronger predictor of youth delinquency than family structure or economic variables. Adolescents with greater attachment were less likely to engage in serious delinquency, including violent behavior. Monitoring of adolescent behavior and friendships has consistently been linked with reduced levels of delinquency (Griffin et al., 2000;Laird, Criss, Pettit, Dodge, & Bates, 2008;Larzelere & Patterson, 1990;Warr, 2005), and to some extent, violence (Windle et al., 2010). Research suggests that parental monitoring may mediate the relationship between attachment and peer formation (Ingram, Patchin, Huebner, McCluskey, & Bynum, 2007;Warr, 2005). Children with strong attachments to parents may be more likely to open themselves up to being monitored (Stattin & Kerr, 2000). They may be more likely to disclose to their parents who they will be with and where they are going.Some researchers have argued that parental supervision directly impacts deviant behavior. Children who are aware that they are being monitored are less likely to engage in deviant behavior (Griffin et al., 2000), while children who are not consistently monitored are more likely to engage in delinquent behavior (Dishion & McMahon, 1998). However, other studies of parental monitoring have suggested that the relationship between parental supervision and delinquent outcomes may be mediated by peer affiliations (Ingram et al., 2007;Laird et al., 2008;Warr, 2005). Warr (2005) finds that parental supervision affects the type of friends that adolescents acquire (see Figure 1). Parents who provide monitoring are less likely to have children who associate with deviant peers. However, this evidence for mediation has been provided primarily in White samples (Forgatch & Stoolmiller, 1994;Oxford, Harachi, Catalano, & Abbott, 2001;Patterson & Dishion, 1985). Further, more research has focused on the mediating role of parenting on peer selection and adolescent delinquency, while less research has focused on examining this relationship for youth violent behavior. --- Peer Affiliation and Violence Perhaps one of the most salient factors related to violent behavior is the influence of antisocial peers (Hawkins et al., 1992;Herrenkohl et al., 2003). In adolescence, young people begin to establish self-identify and autonomy while starting to distance themselves from parental authority (Muuss, 1996). During this stage of life, peer influences on adolescent thought and behavior increase. Youth commonly adopt peer group norms, values, and behavior, and seek peer support and acceptance. Many studies demonstrate that exposure to negative peers is the most important risk factor for antisocial behavior, including violence (Hawkins et al., 1998;Lipsey & Derzon, 1998;U.S. Department of Health and Human Services, 2001). If Black youth are exposed to more negative peer influences, then negative peers may mediate the relationship between race and violent behavior (Haynie & Payne, 2006). While the role of prosocial peers has been less studied there is some evidence that strong bonds with prosocial peers can buffer or mitigate the impact of negative peer influences on problem behaviors, including violent behavior (Kaufmann, Wyman, Forbes-Jones, & Barry, 2007;Kobus, 2003). This study examines the relationships between positive parenting practices and deviant peers and the frequency of self-reported violent behavior in the 10th grade. We also examine race differences in mean levels and the impact of these risk and protective factors on selfreported violent behavior. The level and impact of family and peer factors on violent behavior across race are modeled prospectively from 8th to 10th grade in a sample of 331 (n Black = 162, n White = 168) families from Seattle, WA using data from self-administered computer-assisted questionnaires. Understanding possible race differences in the role of parent and peer risk and protective factors will lead to improvements in targeted preventive efforts in family and school-based programs. Mean-level differences in risk factors can be addressed universally. However, if the risk and protective factors which influence violence are different in each group, preventive efforts must ensure that the appropriate factors are targeted by preventive efforts. --- Methods The current study draws on data from a longitudinal study of 331 families, half who are White, and half who are Black, also balanced by gender. The longitudinal study includes an experimental trial of the impact of Parents Who Care, a preventive intervention to reduce problem behaviors. Previously published reports on the effects of the intervention found reduced frequency of violent behavior for Black, but not White teens (Haggerty, Skinner, MacKenzie, & Catalano, 2007). To control for possible intervention differences in the impact of parent and peer predictors of violence, intervention group was included as a dummy variable in the multiple-group structural equation model (MGSEM). These dummy variables were not significant in these models, and MGSEM by race found no significant differences by race in the impact of intervention group on the outcome or the mediating effects of associating with deviant peers, or on the magnitude of relationships in the models tested. --- Study Sample, Design, and Setting Parents of eighth-grade students in Seattle Public Schools received a letter describing the study, and the parents were contacted by phone. Families were included if the teen and one or both parents consented to participate. Eligibility included self-identifying as Black or White, speaking English as their primary language, and planning to live in the area for at least 6 months. Forty-six percent consented (55% of Blacks and 40% of Whites). The parents who refused were more likely to be White, married, and had a higher education on average than those who consented. Three hundred and thirty-one (n Black = 162, n White = 168) families consented to participate. All protocols were approved by the University of Washington (Seattle, WA) internal review board. The sample was stratified by teen race and gender to meet the requirements of the original randomized trial of Parents Who Care. Sufficient numbers from each group were necessary to test for race differences in the efficacy of the intervention. Families were recruited based on parental report of teen race to the school district (African American or European American) and gender (male or female) into a design with roughly equal numbers in each of the four cells. Recruitment of new families stopped when all four cells were full. There were significant differences by race on several demographic variables. Whites reported higher per capita income and parental education, and Blacks reported higher prevalence of single parenthood (Table 1). Some teens in each race group self-identified as mixed race (19.6% Black; 12.5% White), but were included in these analyses because the study design was based on parental report to the school district and because teen selfreported race changed somewhat over time making categorization more difficult. Most primary caregivers were female (> 80%), with 71.6% being the adolescent's biological mother. Gender and relationship of caregiver were similar across race with one exception: more Black youth had another female caregiver (e.g., grandmother, aunt) as a primary caregiver than did White youth [X 2 (1) = 13.95, p <unk>.001]. Data collectors went to the families' homes. Questionnaires were self-administered to teens and their parents in their homes using laptop computers while the data collector was present. Efforts were made to have the questionnaires completed privately and parents did not monitor their teens' responses. Identical questionnaires were collected when the teens were in the 8th, 9th, and 10th grades. Family members received $15 each time they completed a questionnaire. Attrition analyses-At posttest, 94.7% of the sample participated in the survey (Black 93.2%, White 95.2%). During the 1-year follow-up, 92.5% of families were interviewed (Black 92%, White 92.9%). Two-year follow-up yielded 92% of the sample, with no completion differences by race. Both parent and teen interviews were completed for 303 families. There were no significant differences between attriters and nonattriters on any of the variables included in these analyses. --- Measures Data on demographics and parenting were collected in the eighth grade. Parents reported their child's race on school enrollment forms. Household per capita income was calculated from parent's endorsement of 1 of 11 categories of annual household income (before taxes). We assigned the midpoint of the range and then divided by the number of people in the household. To reduce the effects of outliers we used a log transformation. Parental guidelines (rules and consequences for substance use) were assessed with six items (alpha =.79) on a 4-point scale measuring parental agreement with statements such as "I have clear and specific rules about my teen's use of tobacco, alcohol, and illegal drugs." Monitoring was assessed with the mean of teen responses to seven items (alpha =.77) on a 4-point scale measuring their agreement, for example, of whether the teen believes his/her parent knows who his/her friends are, where the teen is, and what he/she is doing. This measure reflect the teen's perception of how much knowledge their parent has about where they are and who they are with, not active attempts by parents to obtain that knowledge. Consequences was measured using the mean of two items (r =.47) on the same 4-point scale: "If you skipped school, would you get caught and punished?" and "If you drank beer or wine without your parent's permission, would you get caught and punished?" Parent-teen attachment was measured using the sum of 28 items (alpha =.94) from the teen report on the Inventory of Parent and Peer Attachment on a 5-point scale from "never or almost never true" to "always or almost always true" (Armsden & Greenberg, 1987). Example items include "My parents understand me," "I trust my parents," and "My parents sense when I'm getting upset." Delinquent behavior and substance use of peers was measured with teen report in ninth grade. The teens were asked to name their three best (or closest) friends (first names or initials only), and were then asked a series of questions about each of those friends: alcohol and marijuana use, getting in serious trouble at school, or having done anything in the last year that could have gotten them in trouble with the police. A dichotomous score was created for each question, with 1 indicating at least one of the friends had engaged in the behavior. Each of the four dichotomous variables was used as an indicator of a latent construct. Frequency of violent behaviors in the last 30 days was measured identically in the 8th and 10th grades. Items included starting a fight, hitting someone to hurt them, hitting parents (not playing), carrying a gun to school, and throwing rocks at people or cars. Summing across the categorical items produced the following categories: 0 = never, 1 = 1 to 2 times, 2 = 3 to 5 times, 3 = 6 or more times. --- Analysis Simple mean differences in level of risk and protective factors were estimated using t-test or chi-square test. Correlations were run among all of the variables and violence. Tests of the theoretical model which examines a model of parental and peer influences on violent behavior were conducted using structural equation models. These models are particularly suited to tests of somewhat complex models of meditational processes and comparisons between groups. They have the added advantage of improving statistical power and reducing the impact of measurement error by modeling latent constructs based on multiple related measures intended to converge on a true score (see Bentler, 1980). Correlated measures of parental influences (monitoring, discipline, attachment, and guidelines) and deviant peer behaviors (marijuana and alcohol use, delinquent behavior, and trouble at school) were modeled as latent constructs (Jöreskog, 1971). These latent constructs have been used in previous reports from this study (Skinner, Haggerty, & Catalano, 2009) and others (Forgatch, Patterson, Degarmo, & Beldavs, 2009;Herrenkohl, Hill, Hawkins, Chung, & Nagin, 2006) as a measure of skilled parenting. Multiple-group confirmatory factor analysis (MGCFA) (Bontempo & Hofer, 2006;Muthén, 1989) was conducted on the two latent constructs: parenting and deviant peers. This step is necessary to confirm that the measures for parenting and deviant peers are comparable for Blacks and Whites, and therefore comparisons in the predictive relationships between constructs are appropriate. Multigroup structural equation models (MGSEM) by race were used to estimate race moderation of the impact of parental influences in the 8th grade and association with deviant peers in the 9th grade on the frequency of violent behavior in the 10th grade, controlling for previous violent behavior, gender, and family income. Parenting and deviant peers were modeled as latent constructs based on the model described above. Goodness of fit indices are compared between a model in which the predictive relationships are constrained to be equal in both groups (Blacks and Whites in this case) and fit indices when all of the predictive relationships are free to vary between groups. A significant difference in the two fit indices suggests that the 'free' model fits the data better than the 'constrained' model and therefore there are some significant differences between parameters in the Black compared to the White sample. If the differences between the fit of the two models is not significant we can conclude that the predictive relationships are the same for Blacks and Whites. A set of ordinal logistic regressions was estimated to test for possible mediation of race differences in violent behavior through the risk factors for which race differences in mean levels were significant. Ordinal logistic regressions are appropriate for testing the effects of multiple factors on an outcome measured in unevenly spaced levels. Our measure of selfreported violent behavior clearly indicates high levels for more frequent violent acts, but is not precise enough to distinguish exact levels. --- Results Group differences in outcomes and predictors are presented in Table 1, arranged by grade. Categorical differences indicated more frequent violent behavior in the 8th and 10th grade among Black than White teens (d =.30). The first question we address is whether there are mean-level differences in risk and protective factors by race. Whites have higher per capita incomes. Black teens experienced stronger family guidelines. Black teens were not more likely to associate with drinking, smoking, or delinquent peers, but were more likely than White teens to associate with peers who experienced serious school trouble. Reports of discipline, monitoring, or attachment to their parents were not different between White and Black teens. The second question examined whether there are race differences in vulnerability to risk and protective processes. First, we examined bivariate correlations separately for Black and White teens. Low attachment to parents was significantly related to violence for Whites (r = -.21), but not for Blacks (r = -.04). The strongest, and only other significant correlation for both Blacks (r =.22) and Whites (r =.19) to violent behavior is association with peers that have trouble at school. However, a variety of differences emerged between a number of the predictors for Blacks and Whites (e.g., peer trouble at school was significantly positively associated with peer alcohol (r =.37) and marijuana (r =.41) use for Whites but not for Blacks (r =.13 and.14, respectively), suggesting that the relationship between predictors may be meaningfully different for Blacks and Whites. We then conducted a multigroup confirmatory factor analysis to verify equality of measurement in the two latent variables by race and gender. The intercept and loading for guidelines were significantly higher in the Black than in the White sample, and were freely estimated, while all other intercepts and loadings were constrained to be equal across groups (see Table 2). These results suggest an acceptable level of equality of the measures across race. Similar analyses confirmed no differences in measurement parameters by gender. In light of measurement equality, the structural model represented in Figure 1 was examined for race differences in predictive relationships. Rectangles indicate measured variables. Circles indicate latent variables. Arrows indicate the predictive relationships between variables. Factor loadings and intercepts were estimated simultaneously with causal paths but are provided in Table 2 rather than included in the figure for the sake of clarity and simplicity. The fully constrained model fit the data (X 2 (56) = 77.82, p =.03, CFI =.958, TLI =.956, RMSEA =.05, WRMR = 1.23 <unk> = X 2 (12) = 5.2, p =.95), and the difference between the constrained and unconstrained models was not significant. Thus, the MGSEM indicated no race differences in predictors of teen violence. Unstandardized coefficients indicated that several of the hypothesized relationships were significant across groups. Males were more frequently violent. Income was significantly related to concurrent parenting in 8th grade, but not directly predictive of violent behavior in either 8th or 10th grade. There was a negative association between positive parenting and violent behavior at eighth grade (double headed arrow). Parenting factors significantly reduced associating with deviant peers one year later. Association with deviant peers increased violent behavior one year later. The impact of parenting on violent behavior was mediated through decreased association with deviant peers in both race groups. Prior violent behavior was a significant predictor of 10th-grade violence. The model explained slightly more variance in violence for Whites than for Blacks (r 2 =.30 vs..24). Finally, a set of ordinal logistic regressions was estimated to test for possible mediation of race differences in violent behavior through the risk factors for which race differences in mean levels were significant (parent education, single parent status, family income, parent report of guidelines, and teen report of friends who get in serious school trouble). Gender and eighth-grade violent behavior were included as control variables in each equation. Parent education, single-parent status, and parent guidelines did not significantly predict violence and therefore did not meet the requirements for mediation, and therefore could not explain race differences in violent behavior. Family income approached significance in predicting violence (b = -.27, p =.058), but when race was included neither race nor income were significant, suggesting that both race and income explain frequency of violent behavior, but the variance they explain is the same, neither contributing uniquely to explain its variability. Having friends who get in serious trouble in school in the 9th grade significantly predicts 10th-grade violence, controlling for gender and earlier violent behavior (b =.88, p =.008). When included with race as a predictor, race was not significant (b =.60, p =.06). In a similar test, with income rather than race included as a predictor, income was also not significant (-.20, p =.174). --- Discussion Violence remains a serious problem that is not equally prevalent among Black and White adolescents. Understanding the mechanisms behind the differences in violent behavior between racial groups can allow us to more clearly specify prevention interventions. This study explored two mechanisms for the differences in violence perpetration between the two groups. First, are there race differences in in exposure to risk factors? We found that Blacks are exposed to higher levels of risk factors such as low income and lower parent education. Black teens experienced more poverty, but also stronger family guidelines. Black teens were more likely to associate with peers with serious school trouble but not more likely to associate with peers involved with alcohol, marijuana, or delinquency. While Black parents report significantly higher levels of family guidelines, there were no differences in parental consequences, monitoring, or attachment to parents. There were, however, level differences in self-reported violent behavior when study participants were in eighth grade (Blacks reported higher frequency of violent behavior than Whites), suggesting the causal process of predictors tested here might be operating at an earlier time point for Black youth than for Whites. The second question explored whether some risk factors are more salient in one racial group than the other. The MGCFA indicated that parent guidelines against substance use significantly loaded on parenting for Blacks but not for Whites. The structural model indicates that in addition to being male and prior violent behavior, parenting affected violence through its influence on ninth-grade peers. Interestingly, analyses suggest that income impacts violence through its positive relationship with parenting. Other research also suggests that differences in violent behaviors are attributable to family contexts after controlling for income inequality (Conger et al., 1992;Heimer, 1997). No race moderation was found, indicating that both Blacks and Whites are equally vulnerable to the effects of deviant peers on violent behavior. We can conclude that race differences in violence at 10th grade are not attributable to lower exposure to skilled parenting or higher exposure to deviant peers since we don't observe mean-level differences in either latent construct. Nor can we attribute the race difference in violence to greater vulnerability to parental or peer factors since race did not moderate the relationships between predictors and violent behavior. The models explain only between 24% and 30% of the variance in violent behavior. What can we make of the racial differences in violence that are not explained by parental and peer differences? We conducted a final set of analyses to further explore what might explain the race difference. We examined the five potential explanations based on race differences in mean levels of demographic and predictor variables. We found that income and race overlap in their explanation of variability in violent behavior to the extent that neither is a unique predictor in this study. Further, both race and income differences were potentially mediated by associating with friends who get in trouble at school (this is the one measure of deviant peers for which there were mean differences by race). The race effect still had a trend level (p <unk>.10) association with violent behavior after controlling for friends who get in serious trouble at school, so caution interpreting this as full mediation for race is warranted. Perhaps this association is only partially mediated. However, the full mediation confirmed for income effects, and the fact that income and race effects are not clearly distinguishable in this sample adds support to the mediation hypothesis for both race and income. Obviously other factors, which we have not explored in these analyses, should be considered. Likely candidates are other individual, social, and environment factors that may be different for Black and White adolescents. Examples of the former are educational experiences, attachment to school and teachers or academic success, or religious involvement. There is reason to suspect educational differences (Hirschi & Hindelang, 1977). We though, expect that race-based differences in environments offer more fruitful opportunities to explain why Black youth are more involved with violence than Whites after parenting and peers have been taken into account. Why do we think this? Even though we have controlled for income here, Black adolescents more frequently live in "challenged" neighborhoods than do their White counterparts and this difference is not subsumed by income. There are both structural and cultural differences in these communities that make them qualitatively different beyond what is measured by family income. For example, Black communities are generally characterized by higher levels of social and economic disadvantage (Sampson, Morenoff, & Raudenbush, 2005), more labor market disadvantage (Crutchfield, 1989), and more of the forced migration produced by mass imprisonment out of and back into these neighborhoods (Clear, 2007). Culturally, the places where Black youths live are more likely to be characterized by the codes of the streets (Anderson, 1999) that emerge when residents experience prolonged profound disadvantage which in turn may explain higher exposure to friends who get in serious trouble at school. Fundamentally, American cities remain highly segregated (Massey & Denton, 1993), and we believe that the next steps towards expanding our understanding of observed racial difference in violence should begin with this essential feature of our social environment. These findings suggest several implications for prevention. Simple analyses of income and race suggest that improving employment opportunity structures for Black parents should reduce the violence disparity. Further articulation of this finding through the structural model suggests that higher income may reduce violence through making it more likely that parents can accomplish good parenting practices and have time for bonding and monitoring their adolescent's activities. Further, our results suggest that programs that promote good parenting practices, including setting guidelines and providing consistent moderate monitoring and consequences and strong teen-parent bonds are likely to reduce violence through the impact on the type of peers with whom teens associate. (Haggerty et al., 2007;Windle et al., 2010). Finally, prevention programs that reduce the number of deviant peers or decrease the likelihood that teens associate with peers engaged in problem behaviors are likely to reduce teen violence (Botvin, Mihalic, & Grotpeter, 1998;Sussman, Rohrbach, & Mihalic, 2004). The strengths of this study include using data from parent and teen reports, the longitudinal design, and the sample of equal numbers of Black and White families. Weaknesses include measurement of each construct at particular ages rather measures of all variables at all age points. Further, missing data were due to attrition over time, but also due to excluding teens who reported having no close friends in the ninth grade. This makes it impossible to say whether having deviant peers is better or worse than having no close friends at all. A final concern is that we did not examine neighborhood contexts which some suggest are important for understanding racial differences in violent behaviors (Sampson et al., 2005). Research suggests that Blacks are more likely to live in neighborhoods characterized by higher rates of poverty and neighborhood violence (Jargowsky, 1996;Massey, 1990). This sample comes from a Northwest city and caution should be taken in generalizing the findings to other geographic areas with different patterns of neighborhood racial disproportionateness. In conclusion, the difference in self-reported violent behavior between Black and White 10th graders appears to be more related to income disparity than racial status. The impact of income on violence appears to be mediated by racially common processes of good parenting practices negatively impacting deviant peer association. The findings presented here shed light on the important contribution of parenting during middle adolescence on the selection of peers in early high school for both Black and White teens, and the direct impact of deviant peers on youth violence. This study adds information that this process applies to the prediction of violence, not just general delinquency as has been found by others, and that this process applies to both Blacks and Whites. Parenting and peer influences on violent behavior from 8th to 10th grade.

This study examines the role that parenting and deviant peers plays on frequency of self-reported violent behavior in the 10th grade, while testing race differences in mean levels and impact of these risk and protective factors. The level and impact of family and peer factors on violent behavior across race are modeled prospectively from 8th to 10th grade in a sample of 331 (n Black = 162, n White = 168) families from Seattle, WA using data from self-administered computerassisted questionnaires. Mean-level differences indicated greater levels of violent behavior and risk for Black teens in some cases and higher protection in others. Multiple-group structural equation modeling indicated no race differences in predictors of teen violence. Income was also predictive of violent behavior but analyses including both income and race indicated their relationships to violence overlapped so neither was uniquely predictive. Subsequent logistic regressions revealed that both race and income differences in violent behavior were mediated by association with friends who get in serious trouble at school. We conclude higher rates of selfreported violent behavior by Blacks compared to Whites are attributable to lower family income and higher rates of associating with deviant peers at school.

At the level of any economic entity, in the decision-making process it is necessary to know and understand all relevant information that shows the links between the principles and pillars of sustainable development, to apply the concept of sustainable development in order to improve and support a healthy economy, an ecological and social system, to follow the integration of social, environmental and economic models in the organisation's strategy, and to prioritise approaches based on the pillars of sustainable development over financial approaches, as these challenges are always beneficial for people and society in general ( Within the 17 Sustainable Development Goals (SDGs) of the 2030 Agenda for Sustainable Development adopted by the United Nations (2015) (UN General Assembly, 2015) addresses complex challenges facing humanity through the framework of the five Ps: people, planet, prosperity, peace and partnerships. Daly (1992) argues that social sustainability includes notions of equity, empowerment, accessibility, participation, cultural identity and institutional stability. Key sustainability factors can be considered as: social capital, social infrastructure, social justice -equity, committed governance (Cuthill, 2009). In a study related to the mixed composition in the boards of directors of companies, when we refer to gens, they led to better quality sustainability reports (Al-Shaer & Zaman, 2016). Sustainability has become a significant alternative to today's business, as the business environment is changing rapidly, oriented and directed by social, economic and ecological developments, companies are thus increasingly interested in corporate sustainability, and through decisions taken at the strategic level, it could be observed the inclusion of the sustainability in the strategy, vision and culture of an organization, by creating a framework to improve social sustainability practices, there being also increasing pressure for the performance of social sustainability, businesses being increasingly analyzed in terms of their social impact ( Eizenberg & Jabareen (2017) elaborate a comprehensive conceptual framework of social sustainability, which is composed of four interconnected concepts of social practices, and each concept has a distinctive function that encompasses major socially oriented aspects: equity, security, Eco-prosumption, Urban forms. In the researchers conducted by Tosun & Leininger (2017) and Olmsted (2021) it is argued that in ensuring social sustainability, it is necessary to build a system that focuses on recognition, reduction, redistribution, reinforcement and reward (5 Rs) to promote gender equality. --- III. RESEARCH DESIGN Sustainability assessment is not only about the environment, it also refers to the assessment of economic and social aspects, the social dimension being an integral part of the concept of sustainability (Janker & Mann, 2020). Over time, we observe a particular interest for all stakeholders in sustainability issues in general and in the concepts of "social sustainability" and "audit quality", in that some stakeholders are directly or indirectly related to the quality of the audit report (AL-Qatamin & Salleh, 2020). Based on the research themes in the literature under review, the authors used VOSviewer software to perform a bibliometric analysis on the terms "social sustainability" and "transparency in auditing". Starting from the literature papers published on the Web of Science platform, for the period 2020-2021, an analysis was performed in terms of the frequency of keywords in the titles and abstracts of these papers, considering relevant a minimum number of 4 occurrences of a word, and out of the 602 keywords, the VOSviewer software reported 32 terms that referred to the established threshold, which led to a composition of the items on 6 clusters according to the resulting map in Figure 1. research (2020-2021) Source: own processing using VOSviewer software According to the map outlined in Figure 1, the six clusters (cluster 1 -red; cluster 2 -green; cluster 3blue; cluster 4 -light green; cluster 5 -purple; cluster 6 -light blue) are visibly correlated with the research topic studied. It can thus be seen that each cluster contains a certain number of items, the first cluster having a number of 9 items, the second cluster forming a grouping of 8 terms, cluster 3 consisting of 5 terms, a number of 4 items forming cluster 4, and clusters 5 and 6 each consisting of 3 items. The content of each cluster is shown in Table 1. --- Table 1. Terms found in the content of each cluster --- Clusters formed Terms grouped in each cluster Red cluster accountability, boundary, corporate governance, decision, good corporate governance, Indonesia, presence, sustainability, sustainability reporting Green cluster assurance statement, big 4, choice, corruption, country, financial auditor, standard, sustainability report Blue cluster content analysis, corporate environmental, environmental information, environmental reporting, quantity Light green cluster 4 csr policy, evidence, nature, supermarket Purpule cluster business model, enterprise, moral economy Light blue cluster Estonia, interest, social audit Source: own processing using VOSviewer software Following the analysis carried out, it can be appreciated that in the studies found in the literature dealing with works on social sustainability and transparency in auditing, the terms that are found interrelated are those related to corporate governance, sustainable reporting, financial auditor, social audit, CSR policy, etc. In the period crossed by the Covid 19 pandemic, many organisations suffered, and this situation disrupted sustainability plans, making sustainability a paradigm in this respect (Bezerra, Gohr and Morioka, 2020). Lately, we find in the literature studies that contribute to the social dimension of corporate sustainability by creating an integrated sustainability framework that includes new approaches in different contexts (Olmsted, --- IV. THE IMPORTANCE OF THE SOCIAL DIMENSION ON THE SUSTAINABLE DEVELOPMENT OF --- ORGANIZATIONS Sustainable development is seen in two senses: an "analytical" one, whereby societies and organizations are given social structures, processes and patterns, which ensures the preservation of useful and functional structures and a "normative" one, which addresses social issues as positive situations desired by society in general (Littig & Greissler, 2005). Sociological considerations are based on the composition of three major indicators that ensure the social dimension of sustainability: basic needs and quality of life, social justice, social coherence (Littig & Greissler, 2005) and other three approaches: "development sustainability", "bridge sustainability", ensure that structures are modified to meet changing needs and "maintenance sustainability", which ensures the preservation of useful and functional structures (Vallance et al., 2011). In the opinion of Gray (2010) social sustainability also encompasses many aspects such as: human rights, gender equity and equality, public participation and the rule of law, all of which promoting peace and social stability for sustainable development and Murphy (2012) identifies four general social concepts that connect with environmental imperatives: public awareness, equity, participation and social cohesion (Murphy, 2012). A sustainable development model has at its core moral desiderata related to meeting human needs, ensuring social equity and respecting environmental limits (Holden et al., 2016) and acording to other authors, the objectives of sustainable development are poverty, illiteracy, access to participation in decision-making (Vavik & Keitsch, 2010). UN Commission for Sustainable Development, UNCSD (1996) presents a series of social categories in the social classifications, namely: combating poverty, sustainable demographic dynamics, protecting human health, promoting human settlements, promoting education, public awareness and training. United Nations Department of Social and Economic Affairs, UNDESA, (2001) relates to the following social issues: equity, health, education, housing, security (crime fighting), population. UN Commission for Sustainable Development, UNDESA (2007) presents as themes in social classifications: poverty, governance, health, education, demography; EU Sustainable Development Indicators, Eurostat (2007) presents also topics related to issues related to: social inclusion, public health, demography, good governance. OECD Social Indicators (2009) (OECD, 2009) considers the social "organizational dimension" expressed by the following indicators: economic self-sufficiency, equity, health, social cohesion. Magis & Shinn (2009) define four universal principles that cover social sustainability: human well-being, equity, democratic governance and democratic civil society. From all the aspects presented above, it can be seen that the social dimension is an essential pillar of sustainable development. --- V. ENSURING THE TRANSPARENCY OF SUSTAINABLE REPORTS THROUGH QUALITY IN AUDIT Conflicts of interest can arise between producers of accounting information and its users, and the credibility of the information can be undermined as third parties want reliable information and managers or directors of companies can gain advantages by providing false or manipulated information. This may be due on the one hand to the fact that information producers are usually not independent of the operations carried out and on the other hand to the fact that the moral and professional qualities of the producers may affect the reality of the information. Thus, based on these situations, an important role will be played by the auditor, who has the task of auditing the financial statements and issuing an independent opinion on the reality and accuracy of the data presented in the annual financial statements. Thus, we note that the general role of financial auditing is to control financial information (the purpose is to protect the organization's assets and ensure the credibility of the information provided to the public) which is of interest both in terms of internal usefulness as a useful support in the management and administration of the organization, and in terms of external usefulness, by informing third parties, mainly represented by the tax authorities, creditors, suppliers, customers, etc. The increase in the relevance and quality of financial information was primarily due to the introduction of IFRS (International Financial Reporting Standards) as the accounting benchmark (Barth et al., 2008). Validation of the information provided by the organization is necessary because its production is the prerogative of management and third parties can only make use of it after the publication of the annual financial statements, which is the final role of the financial audit in verifying and certifying the financial reports. In the specialized literature, there are studies that have shown that changes in audit reporting standards support the development of the audit profession, consolidating and academically completing the credibility offered by the auditor (Bunget et. al., 2012;Czerney et. al., 2014;Bédard et al. et al., 2016). According to its historical development, the usefulness of financial auditing is also justified by the detection of errors, fraud and deception as an important segment of the audit mission in recent years, complementary to the knowledge and promotion of standards, the assessment of the internal control system and providing assurance on the accuracy of the information presented through the financial statements. Improving the integrity and efficiency of financial statements and increasing transparency in financial reporting (Directive 43/2006) is ensured by a quality financial audit (McPhail & Walters, 2009). Audit transparency can be determined on the basis of qualified or unqualified opinions, but with insignificant comments, expressed by the financial auditor (Robu et. al., 2019). The reporting of companies has improved over time, in order to satisfy more and more the growing needs of interested parties, which are oriented towards an integrated reporting, in order to fulfill the transparency and responsibility characteristic of financial and non-financial reporting (Demirel & Erol, 2016). By ensuring the transparency of information in audit reports, their communication value to stakeholders also increases, this improvement in communication being supported by the introduction of KAM in the content of the audit report (Backof, 2015 Gender equality, as a constructive part of the social dimension conceptualized as multifactorial and multidimensional, underlies the different dynamics between women and men in terms of segregation in sectoral and technological contexts, a concept that is seen as a component of the concept of corporate social responsibility through the way specific normative principles are respected (Risman, 2004;Zander et ) that pay attention to the behaviors of those involved in producing, validating, publishing and capitalizing financial statement information, from a gender perspective have shown that female accounting professionals are more risk averse, engage less in unethical behaviour for financial gain, their presence on boards of directors, through their prudent behaviour, influences the reporting of lower results, and their presence on audit committees also leads to quality financial audit engagements, thus requiring higher audit fees. The greater number of women in the accounting and auditing profession is explained by some authors (Dimnik & Felton, 2006) by a number of gender-specific characteristics, namely: patience, prudence, carefulness, thoroughness, good analytical skills, hard work, practicality, sociability and respect for principles and conventions. In some countries, such as for example in Sweden at the beginning of the certified audit profession, there were significant differences in the assessment of women's and men's performance, career intentions and career plans in the audit field (Jonnegard et al., 2010). Other studies (Carmona & Ezzamel, 2016) have shown that in the exercise of activity in the field of accounting and auditing, gender differences can be perpetuated, due to accounting and auditing technologies and the vocabulary used in assessments. The objective of auditing has undergone a number of changes, just as the role of the auditor and his work has incorporated the influences of the economic and social environment specific to its evolutionary stages. To a certain extent, the current objective of the financial auditor incorporates most of those previously assumed (such as punishing those who misappropriate public funds, preventing and combating fraud, providing assurance on the true and fair view of financial statements, and protecting against national and international fraud). --- VI. CONCLUSION A preliminary conclusion on the analyzed aspects leads us to the idea that the sustainable performance of organizations by applying audit tools represents topics of interest in the scientific body of research, even if the search on the Web of Science platform did not generate many results in this direction. To ensure social sustainability, gender equality is a key objective through ethical and fair corporate practices. This aspect has influenced research from various fields, such as accounting, auditing, finance and management, which have taken into account the gender variable, to explain some behaviors of actors involved in the production, validation, publication and exploitation of information from financial and non-financial reports. Ensuring the sustainable performance of an organization is not only related to financial performance, but also aims at the social and environmental dimension, systematically framed in the concept of global performance, which involves access to resources for exploitation, redistribution in order to obtain cash flows to eliminate potential risks and in order to achieve the objectives. In the context of social sustainability and transparency of audit information, we consider that the results obtained through bibliometric analysis bring to the fore a detailed correlation of these concepts with the issue of sustainability and sustainable performance of organizations, this situation leading to a solid scientific basis for future empirical research.

Social, economic and environmental factors contribute to the sustainable development of organizations. The social and financial dimensions have an important contribution to sustainable performance, through the insurance of transparency in the communication of information requested by stakeholders in order to substantiate their decisions. Social transparency is ensured by presenting the organization's actions in the field of social responsibility, and financial transparency takes into account the most accurate, complete and neutral presentation of information, both in the financial statements and in the audit reports. The main objective of the paper is the analysis of the opinions regarding the contribution of the financial audit to the sustainable development of organizations through social transparency.

INTRODUCTION The current model of funding of higher education in South Africa relies upon government subsidies to higher education institutions (HEIs); third stream income generated by HEIs and tuition fees paid by students to HEIs (Salmi and D'Addio 2021, 57). In 1999 the government introduced the National Student Financial Aid Scheme (NSFAS) to extend opportunities for fee paying students from disadvantaged backgrounds to access higher education. The NSFAS is funded by the Department of Higher Education and Training (DHET) and is operationalised within the National Student Financial Aid Scheme Act 56 of 1999 in tandem with the Skills Development Act 97 of 1998. The core objective of the NSFAS is to facilitate equitable access to higher education by providing financial support to eligible students (Matukane and Bronkhorst 2017, 3). The NSFAS currently operates at all South African HEIs and provides eligible students with opportunities to access loans or bursaries to enable them to pay their fees and support themselves financially whilst studying. The NSFAS currently offers loans to students who provide evidence that their circumstances are such that financial assistance is warranted and have demonstrated the potential to complete their studies (Wildschut, Rogan, and Mncwango 2020, 964). The NSFAS also provides scholarships -comprising a maximum of 40 per cent of the value of the loan -to students who successfully complete their studies within the minimum allowed period for the qualification loan (Wangenge-Ouma and Cloete 2008, 907). Since its inception in 1999 the NSFAS annual funding allocation increased from R441 million to almost R42 billion in 2021 (NSFAS 2021, 4). The scheme is inclusive of all registered students at both Universities and Technical and Vocational Education and Training (TVET) colleges. With the COVID-19 pandemic and subsequent national lock-downs, HEIs, and education at large, were affected at multiple scales (Cesco et al. 2021, 276). The broader economic constraints associated with the pandemic affected the South African higher education budget and subsequent allocations to HEIs. The NSFAS funding operational strategy was also affected -which in turn catalysed policy reviews and budgetary re-allocations (Schreiber et al. 2021, 15). --- PROBLEM STATEMENT Whilst the ambition of providing youth with equitable opportunities to access HEIs through the NSFAS was both a coherent and relevant framework which was designed to boost the South African skills base, there have been critiques with regard to implementation. The critiques revolve around issues such as the inclusivity of the scheme (Garrod and Wildschut 2021, 488); insufficient consistent communication between the administrators of the scheme and the student beneficiaries (McKay, Naidoo, and Simpson 2018, 25-27); inadequate levels of transparency between the administrators of the scheme and the student beneficiaries (Manuel 2019, 77) and operational bottlenecks with regard to timely payments and repayment inconsistencies (Yende 2021, 73-76). Notwithstanding these critiques, NSFAS's mandate remains one of delivering opportunities for students to both access higher education and "improv[ing] the quality of education, skills development and innovation" of them whilst they are studying (NSFAS 2020, 13). Currently, the relevance of NSFAS is especially underpinned by the National Development Plan 2030 (NDP) which explicitly makes the case for increasing the number of skilled graduates as a contribution towards accelerating national growth and reducing the disproportionately high levels of unemployment in the country (Balwanz and Ngcwangu 2016, 32). The relevance of the scheme is also reinforced by efforts to place sub-Saharan Africa as an internationally competitive region within the global economy by developing "skills for the workforce of the future" (Ocampo and Ortega 2022, 235). Despite the strategic focus on skills development at both international and national levels, research into the efficacy of the NSFAS contribution to skills development in South Africa is scant. Whilst peer reviewed literature relating to skills development in South Africa itself is relatively abundant (for example, a Google Scholar search for "Skills development South Africa" reveals 31 600 hits since 2018); there is a relative scarcity of the same for NSFAS and skills development (a Google Scholar search for "Skills development South Africa NSFAS" produced 1 680 hits within the same period). Even more scarce is literature relating to the perspectives of the student beneficiaries of NSFAS support and their experiences regarding their acquisition of skills whilst studying. The study thus represents an embryonic, qualitative step towards providing insights from students about their experiences relating to skills development whilst being beneficiaries of NSFAS support at the University of Limpopo. The modest contribution made in this article does provide new insights to the existing literature relating to NSFAS and skills development from the perspective of beneficiaries which may also provide insights for policy makers going forward. --- HIGHER EDUCATION: SOUTH AFRICA During apartheid access to higher education was minimal among the African population. With the shift into a democratic dispensation South Africa embarked on a radical educational policy shift which was designed to enable equitable access to education for all (Boughey and McKenna 2021, 13). The White Paper of 1997 on Higher Education also identified four pillars of higher education: "1) To meet the learning needs of individuals in order to achieve self-fulfilment; 2) To provide the labour market with highly skilled graduates; 3) To contribute to the development of a critical citizenry and lastly to contribute to knowledge creation" (White Paper on Social Welfare 2014, 34). Since the commitment to boosting skills through education in South Africa emerged almost 25 years ago, there has been a rich debate about the types of skills that are relevant for students to acquire. Typical, this discussion categorises skills into two broad themes: hard and soft skills. Hard skills are typically associated with technical skills, such as engineering and construction (Maisiri, Darwish, and Van Dyk 2019, 94); whilst soft skills are behavioural, such as interpersonal and communication (Assan and Nalutaaya 2018, 62). It has also been argued that the acquisition of soft skills is premised upon "learning by doing", rather than qualifications alone (Lyu and Liu 2021, 15). Alongside the experiential learning associated with soft skills, it is also suggested that both individual and group commitment to a task; attitudes towards a task and determined confidence to complete a task are also mediators of soft skills development (Succi and Wieandt 2019, 117). Not only do soft skills represent a portfolio of attributes, soft skills are also believed to add value to technical hard skills which tend to be habitual, rather than because of "learning by doing" within different particular contexts (Majid et al. 2019, 20). Whilst hard skills are focused on in more depth at South African HEIs, many argue that soft skills are equally important component of curricula, as well as providing broader life opportunities for students (Pezer 2021, 62;Boughey and McKenna 2021) -and Ngulube (2020, 1002) has argued that, in some instances, "[South African] employers are more interested in soft skills than hard skills". --- SKILLS DEVELOPMENT AND EMPLOYMENT IN A SUB-SAHARAN CONTEXT It is well known that unemployment in sub-Saharan Africa has, and continues to be, -a significant challenge in multiple ways. Recent evidence suggests that approximately 60 per cent of the unemployed in sub-Saharan Africa are younger than 25 and that young women are disproportionately affected (Abodohoui and Su 2020, 10). In response to this challenge there have been multiple interventions aimed at reducing the disproportionally increasing rate of unemployment amongst the youth in sub-Saharan Africa (Cieslik, Barford andVira 2022, 1138). (Mseleku 2022, 330). --- SKILLS DEVELOPMENT AND NSFAS The South African Skills Development Act 97 of 1998 (SDA) was designed to integrate the mandate of the National Skills Authority with institutions that provide learnerships and/or skills development opportunities for South Africans -as well as mechanisms to provide financing for both in cooperation with the National Skills Fund (NSF) (Plant and Padotan 2017, 37). The initiative is underwritten by the necessity of ensuring sustained investments in both learnerships and skills development opportunities as a contribution towards boosting national growth, reducing unemployment and increasing household prosperity in an equitable manner across the nation. A further contribution to the mandate of the SDA is made by HEIs which proactively seek out opportunities for students to gain experience in a work environment. These opportunities are manifest in research, community engagement and work integrated learning (WIL) projects and programmes. In tandem with the national focus on providing experiential training opportunities for youth to become prepared for the world of work, the government has introduced funding mechanisms at HEIs to enable marginalised youth to access higher education. This is because whilst the higher education system in South Africa has always had a fee-paying component and HEIs generate third stream income in multiple ways, a shortfall in funding for HEIs is a national reality, hence additional measures have been taken by government to provide support to financially disadvantaged students (Menon and Cedras 2019, 246). Despite the institutional efforts to facilitate sufficient funds for higher education it became evident during the early years of the democratic dispensation that the available funding could not accommodate all youth who were eligible to access HEIs -especially those from disadvantaged backgrounds. The gap between eligibility to enter an HEI and funding for potential entrants has become particularly noticeable in the last decade with increasing numbers of eligible youth requesting access to higher education who are simultaneously unable pay tuition fees (Ayuk and Koma 2019, 14). Despite the extensive funding problems that have been encountered in South Africa, the state has attempted to implement a combination of both integrated and comprehensive funding to support both higher education and skills development for students (Naidoo and McKay 2018, 160). It is within this context that NSFAS operates. The NSFAS funding framework was initially developed and continues to operate within the context of hard budgetary constraints and is driven by policy reviews of South Africa's spheres of government for higher education, rather than the cost of provision (Menon and Cedras 2019, 248). The NSFAS funding framework developed for HEIs in South Africa has a number of important implications for promoting equity, fairness and recognition of institutional development in the context of education and skills development. Education and skills development are drivers of personal, national and global development, therefore investing in both is essential because they have an impact on South Africa's economic development (Turek 2020, 160). The National Skills Development Strategy (NSDS) was another policy that was initiated in the post-apartheid South Africa as a component of the NSDS to strengthen and advance skills development. The NSDS continues to represent a strategic framework of cooperation between the state, businesses, institutions of higher learning and the labour market (Kraak 2008, 14). In South Africa's HEIs students are motivated to actively participate in practicals, projects and credit bearing WIL initiatives as well as other hands on programmes. These hands-on initiatives span a broad spectrum of fields, including, inter alia -peer education, the automotive industry, electrical engineering alongside student mentorships as a way of promoting skills development through experiential learning opportunities. At the University of Limpopo skills development is reflected through the student support service programme, whereby in the year 2019 the university achieved 60 per cent of the target it had set for improving student support and providing development services (University of Limpopo Annual Report 2019, 42). --- RESEARCH METHODOLOGY The method applied in this study was an exploratory, qualitative field research design. Fieldstudy research can be described as "participant observation or design around a social field which contains a variety of interpretative procedures and interactions studying people individually or in groups while inhabiting their natural environment by using research tools or other forms of gathering data" (Queirós, Faria, and Almeida 2017, 23). The study focused on the lived experiences of students at the University of Limpopo to assess the utility of the NSFAS towards skills development from the perspectives of student beneficiaries. --- Participants The participants for the study were purposefully selected using a non-probability sampling method. Purposive sampling means that the researcher does not attempt to recruit a random sample of participants, but rather selects participants based on their knowledge or familiarity with the phenomenon being studied. This enables researchers to identify a predetermined set of selection criteria and identify the participants accordingly so that the data collection process engages with participants that have the experience required to make a meaningful contribution to the overall aim and objectives of the study (Creswell 2016, 21). The sample was drawn from the Faculties of Management and Law and Science and Agriculture at the University of Limpopo. In total eighteen (18) students from second to fourth levels of study participated across the two faculties. There were three participants per level of study per Faculty. Part of the selection criteria for the study was that the participants had received support from the NSFAS for two to four years. The reason for this was to ensure that the selected participants had insights into the way in which the NSFAS funding had facilitated exposure to skills development opportunities. --- Data collection and analysis The research was undertaken during the latter parts of the COVID19 lock-downs and thus appropriate measures were undertaken before, during and after data collection to ensure compliance with the associated national regulations and guidelines. This included wearing a face mask at all times; use of hand sanitizer; sanitizing the recorder and notebook and maintaining a distance of two metres at minimum between the interviewer and interviewee. These precautionary measures did not affect the outcomes of the interviews in any noticeable way. In addition to the steps taken to comply with the national COVID-19 regulations and guidelines, conventional ethical protocols were adhered to. Table 1 provides a summary of the ethical protocols that were complied with during the data collection phase of the study. A semi-structured interview protocol with individual participants was the primary data collection method, augmented by observations of body language during the interviews. Semistructured interviews were preferred because they allowed for a conversation with a peer (interviewer) which made for a relaxed environment for the participants to express their opinions about their experiences about the utility of NSFAS towards their skills development. The semi-structured interviews were conducted at the University of Limpopo student cafeteria and took a period of one month. The cafeteria provided a neutral, yet convivial, setting for the interviews which enabled the researcher to conduct formal interviews in a relaxed manner that provided the opportunity for clarifications and probing to occur as required. Following the semi-structured interviews, the researcher transcribed the recordings of the eighteen interviews prior to undertaking a thematic analysis of the data. Table 2 provides a summary of the phases of the analysis. The process was iterative and required both the narrative data and initial organisation being reflected on 4. Theme development --- • From the codes, higher level conceptual themes were identified • Once again, this was an iterative process that required re-visiting the material generated in the earlier phases of the analytical process Source: Authors' contribution The analysis followed the steps recommended by Creswell (2016, 30) which included the initial categorisation of the data prior to coding and theme development. During the analysis the NVIVO computer-assisted qualitative analysis software (CAQDAS) was used to manage the data. In order to consolidate the thematic analysis, the researcher critically reviewed and adjusted the themes until a coherent overview emerged. --- FINDINGS The theme refinement process identified five themes which are listed below. --- Financial security; 2. Spending prioritisation and collaborative budgeting; 3. Time-management skills; 4. Self-esteem (self-confidence and personal growth), and 5. Skills development opportunities. Details of the five themes are provided in some detail in the section that follows. Each subsection also contains direct quotes from participants to justify the labels given to each theme. --- Financial security Financial security was a dominant theme that primarily related to the provision of food, followed by domestic items whilst studying at the University of Limpopo. "As soon I received my allowances in a form of direct payment into my student bank account, I bought groceries such as maize meal, meat... and all other necessities." (Interviewee 5, female). Secondary domestic items were also referred to, but less frequently than the provision of food. "The funding allocation helped me to buy food and the rest of the money I would use it for clothes, hair, cosmetics and beauty products." (Interviewee 3, female). The narratives provided by the participants relating to food security and domestic items have some resonance with Sano, Mammen, and Houghten (2021, 110) who identified students having sufficient savings to enable them to procure food and other expenses whilst studying as being priority areas. However, the NSFAS funding not only provided some financial security for students, it also provided them with opportunities to prioritise the items that they spent their funds. --- Spending prioritisation and collaborative budgeting This theme provides indicators of participants' prioritisation of NSFAS funding allocations. "When I received my funding allocations, I would write down a list of what needs to be paid firstly. For example... accommodation, transport and food... and allocate a specific amount needed for each category. In this way, it was simple for me to manage my funding allocations." (Interviewee 2, female). Not only did the participants report that they prioritised their spending allocations, they also reported that they would often collaborate with other students when budgeting. The participants simultaneously spoke about prioritisation of their budget and the collaborative efforts during the interviews which is why the two themes were merged into one. "I stayed with my roommate, so a sharing of funds would occur between us. In this manner we would minimise the expenditure costs and have more money to save." (Interviewee 6, male). This theme resonates with findings from other studies relating to student well-being whilst studying. For example, Azer andMohamad (2018, 2474) report that students who master the skill of budgeting at an early stage are more likely to complete. Likewise, Dalla Pellegrina et al. (2021, 10) argue that the "prioritisation of spending by students is an important skill and the learning process simultaneously enables personal growth and entrepreneurial financial responsibility" which may in turn be the first step towards acquiring a lifelong skills relating to personal financial management (Rosskam 2018, 211). Following the above two themes, the participants also reported that time management was another skill that the NSFAS support enabled. --- Time management skills Due to the reality that the NSFAS funding allocations are released monthly the students reported that they began to manage their time accordingly. Once again, this theme is not unique. For example, Alias et al. (2019, 127) argue that "time management is a skill that is acquired through experience which results in individuals improving how they organise their time effectively in order to maintain productive results whilst studying". In this instance the experiential skill of time management was acquired because of the stability the NSFAS funding framework provided. Arguably, because of the skills acquired by student participants whilst managing their NSFAS funds described above, their self-esteem improved which is the subject of the following theme. --- Self-esteem (self-confidence and personal growth) The notion of'self-esteem' is applied broadly in this instance because it refers to multiple forms of esteem including self-confidence, reduced self-doubt and privilege. Some examples are provided below. "Being able to become a NSFAS beneficiary motivated me as many people get excluded through the NSFAS admission process and it is a privilege to be funded." (Interviewee 8, female). "The funding allocation stimulated my self-confidence and reduced self-doubts. I managed to save my funds and was able to register for a driving licence as most jobs in my career pathway require the proof of having a driving license." (Interviewee 11, male). This theme becomes relevant when other literature is considered. For example, Yu, Shek, and Zhu (2018, 23) articulate that "self-confidence and, to a lesser extent, personal growth affects students' educational goals and may develop into a holistic life-skill". Whilst this suggestion that "self-confidence affects students' educational goals" is hardly surprising, it is relevant that the NSFAS funding enabled the personal growth of marginalised students who may not otherwise have had such an opportunity. --- Skills development opportunities The final theme that emerged related directly to skills development opportunities. Despite the NSFAS funding providing indirect opportunities for skills development, there were no reports that the NSFAS provided direct skills development opportunities. "NSFAS has not provided any opportunity, but through the institution they are various opportunities and seminars that are provided for different faculties, in order to grant exposure of various fields of occupation and where you find interest at." (Interviewee 12, male). "Some funding schemes offer the privilege of finding jobs to their beneficiaries. In the context of NSFAS, I have never encountered any opportunities." (Interviewee 13, female). This finding dovetails with other research within the African sphere such as Darvas et al. (2017, 15) who argue that "the lack of skills development opportunities at African HEIs hinder students' opportunities and development within working environments in preparation for the world of work". --- SOFT SKILLS The themes that were identified during the analysis suggest that the NSFAS support provided to the participants indirectly contributed to their soft skills development. The themes that emerged from the analysis indicate that soft skills were acquired through the secure context the NSFAS funding provided -and the context represented a framework which afforded the participants to experientially learn the soft skills described above in the themes. Other studies have emphasised the relevance of students acquiring both soft and hard skills whilst studying and noted that the combination of both contributes to students' personal and educational growth (Rebele and Pierre 2019, 75). It has also been noted above that soft skills are an attribute that many employers actively seek out when recruiting graduates, yet as Spanjaard, Hall, and Stegemann (2018, 166) argue, HEIs rarely bestow soft skills development with the same degree of focus as hard skills. This debate is summarised by Mtawa, Fongwa, and Wilson-Strydom (2021, 687) who argue that students should be equipped with soft skills for them to become "players in the labour market". --- DISCUSSION The themes revealed in this study indicate that the soft skills described above by the participants who were NSFAS beneficiaries enabled them to identify opportunities whilst studying at the --- University of Limpopo to develop effective mechanisms to support themselves whilst studying. From this perspective, it is suggested that in some instances, the utility of NSFAS funding is an indirect catalyst that contributes towards students being able to develop their own, tailor-made, personal enabling environment that facilitates the space for them to commit to their studies. Likewise, it is plausible -but difficult to provide supporting evidence -that the soft skills that were experientially acquired may also contribute towards their longer-term personal development which may develop further into life-long skills and also represent the early phases of preparing themselves for the world of work. Because of potential relevance of the findingsand despite the localised, exploratory nature of the study -further research into the utility of NSFAS funding to promote soft skills development among students in South Africa is warranted. --- CONCLUSION Skills development is inclusive of people acquiring both hard and soft skills -with HEIs generally focusing efforts towards hard skills development. Hard skills can be understood as skills that are taught through theoretical or formal education and are typically reflected in associated qualifications and soft skills are associated with more abstract traits such as selfesteem. The narratives from the participants suggest that the utility of NSFAS funding for beneficiaries was that they were able to develop soft skills as an indirect consequence of the NSFAS. The indirect effect of the NSFAS support was that it provided the student participants opportunities to engage in experiential learning about personal development whilst managing their funds in ways that made sense to them in their context at the University of Limpopo. The combination of soft skills acquired whilst the students managed NSFAS funds thus contributed to the students being able to commit to their studies and may plausibly represent a set of soft skills that could contribute to preparing them for the world of work. This latter point -being prepared for the world of work -could be an area that HEIs can investigate further and consolidate into a training programme to assist students whilst managing their NSFAS funds, as well as to use the learning process as an opportunity to prepare students for the world of work. --- DECLARATION OF INTERESTS The authors declare that they have no personal or financial interests or relationships that could have influenced this study or its findings. The authors ascribe to the highest standards of ethical conduct in all research endeavours. This study was approved by the University of Limpopo Turfloop Research and Ethics Committee and granted an ethical clearance certificate with the project number: TREC/404/2020: PG.

The National Student Financial Aid Scheme (NSFAS) was introduced in 1999 in South Africa to provide financial support for disadvantaged students through the provision of both loans and bursaries. This article reflects on the effects of the scheme on skills development within two faculties at the University of Limpopo. The research was undertaken in 2021 employing an exploratory qualitative methodology. A total of eighteen participants were purposively selected from two of the four faculties of the University of Limpopo. A semi-structured interview guide was utilised to elicit narrative data from all participating beneficiaries of the NSFAS. The narrative data were analysed using a thematic coding design and interpretations of the findings were subsequently made. The interpretation of the analysis focused on the effect of the NSFAS support on students' skills development. The findings suggest that NSFAS funding contributed to the development of soft skills of beneficiaries which included financial security; spending prioritisation and collaborative budgeting; time-management skills; self-esteem and skills development opportunities. Whilst the findings are localised, they do suggest that further research could be undertaken into the potentials of NSFAS to boost students' soft skills development during their time within a higher education institution.

Introduction Bullying is intentional and repeated aggressive behavior toward another person in which there is a real or perceived power imbalance, and the victim of bullying feels vulnerable and powerless to protect themselves [1][2][3]. Bullying includes physical assault, verbal abuse, and neglect [4]. Globally, bullying is widespread among adolescents. In a 2018 report by UNICEF, more than one-third of students aged 13-15 worldwide said they had experienced different forms of bullying [5]; data published by the World Health Organization in 2020 showed that more than 100 million children worldwide died each year from violence, including severe domestic violence as well as bullying [6]. In a survey involving 40 developing countries, the results showed that an average of 42% of boys and 37% of girls had experienced or were experiencing bullying [7]. Evidence from several longitudinal studies on the effects of bullying suggests that experiencing bullying, especially in adolescence, can severely impair a person's physical, psychological, and social functioning, leading to risky behaviors [8], anxiety [9], depression [3,10], lower levels of academic achievement [11,12], suicidal ideation, suicidal behavior, or self-harm [13][14][15]. At the same time, as a deliberate, repetitive act of aggression that occurs when there is a power imbalance between the perpetrator and the victim, the perpetrator repeats the bullying against the victim, and the repetition tilts the "balance" between the perpetrator and the victim, making it difficult for the victim to escape from the situation [2,4], thus having a lasting psychological effect on the victim [16,17]. This has a long-lasting effect on the victim's psyche. Research has shown that the frequency of bullying is one of the factors that affect adolescent mental health. Adolescents are more likely to experience more severe depression when they are bullied more frequently [4], and some victims of bullying may even become perpetrators of bullying, harming peers or others [18][19][20]. In recent years, some studies have also begun to further explore the effects of different forms of bullying on adolescent mental health, and found that the form of bullying is also an essential factor affecting adolescent mental health. The first was to explore what forms of bullying had a profounder effect on adolescent mental health, but most of the current studies by researchers on this issue have been conducted in individual countries or regions and have not reached uniform conclusions, e.g., Maunder et al. (2010) conducted a survey of students, teachers, and staff in four secondary schools in England, and a total of 1302 people participated in this survey, and the results found that physical bullying was the most harmful to students [21]; Chen et al. (2012) selected a middle school in Taiwan, China, and conducted two samples (605 students and 869 students) and found that relational bullying such as rumor spreading and cyberbullying were more harmful than physical and verbal bullying [22]; Thomas et al. (2016) selected 10,273 secondary school students in the first adolescent health survey conducted in 2009 in Victoria, Australia, and found that neglect had the strongest association with mental health among four forms of bullying (teased or called names, spread rumors, neglect and physical bullying) [4]. In a representative cross-sectional standardized survey conducted by Baier et al. (2018) in a federal state of Germany (10,638 students in the 9th grade), psychological cyberbullying was found to be the most important influence on the mental health of boys and girls, followed by relational bullying from peers or from teachers, and girls' mental health was associated with sexual cyberbullying. There was no significant effect between physical bullying and mental health [23]. The second was to focus on the effect of different forms of bullying on adolescent mental health under the gender group. For example, Turner et al. (2013) selected 1874 students from middle and high schools in North Carolina to explain the results of the effects of different forms of bullying (physical, verbal and cyber) on mental health (including depression and suicidal intention) and found that females had higher levels of depression after cyberbullying compared with males, and there was no significant difference in suicidal intent after either form of bullying for either males or females [24]. Shongwei et al. (2021) used the database of the 2013 Eswatini Global School-based Student Health Survey to examine gender differences in the effects of different forms of bullying on mental health in a sample of 2920 children aged 15-17 years, and found that both boys and girls felt lonely and feared re-victimization after being bullied [25]. Using data from U.S. Youth Risk Behavior Surveillance System in 2015, Kim et al. (2019) found that school bullying had a greater negative psychological effect on girls than on boys [26]. Wang et al. (2009) selected a sample of 7182 U.S. adolescents in grades 6 to 10 based on the 2005 Health Behavior in School-Aged Children Survey, and found that boys were more likely to involve physical or verbal bullying and girls were more likely to be involved in relational bullying [27]. Very few studies have focused on the effects of different forms of bullying on adolescent mental health according to age groups, and Yen et al. (2014) found that adolescents in middle school had more severe mental health problems after bullying than those in high school [28]. In addition to exploring the negative effects of bullying on adolescents, there were very few studies that analyze the role and effect of protective factors in preventing the occurrence of multiple forms of violence as positive actions to build resilience in children, in terms of protective factors [29][30][31]. For example, Biswas et al. (2020) used data from the Global School-based Student Health Survey to divide protective factors into parental support and peer support, and explored the effect of each on the mental health of adolescents following bullying [32]. Although some studies have been conducted on the effect of bullying on adolescent mental health, there are still the following research gaps: Firstly, for the global prevalence of adolescent bullying, the current studies are mostly limited to one country or a few regions [33,34], the findings are not consistent across countries, and there is a lack of crossregional comparative studies. Secondly, in addition to focusing on the effect of bullying on the mental health of adolescents as a whole and different gender groups, there are not enough studies on the effect of bullying on the mental health of different adolescent subgroups. Adolescents are at a critical stage of development and the influence of age on their behaviors is crucial, but there is a lack of research discussing the effect of different forms of bullying on mental health according to age groups. Thirdly, current research has focused more on the risk factors of adolescent bullying and not enough on protective factors [31,35,36]. Therefore, to address these limitations, this study attempts to analyze the frequency of bullying, the prevalence of different forms of bullying, and the effects of both on adolescent mental health in 65 countries from a cross-regional comparative perspective, and to explore the differences in the effects of different forms of bullying on adolescent mental health by gender and age groups in different regions. In addition, the study focused on parental support as a protective factor to examine the relationship between parental support and the mental health of adolescents who experienced bullying, and the mental health of adolescents who experienced different forms of bullying. The following were our hypotheses: Hypothesis 1 (H1). Forms of bullying would be associated with adolescents' psychological well-being. --- Hypothesis 2 (H2). Forms of bullying would have significantly different effects on different genders and ages. --- Hypothesis 3 (H3). Parental support would play a moderating role in psychological well-being of adolescents who experienced bullying. --- Materials and Methods --- Data and Sample Global School-based Student Health Survey (GSHS) is a World Health Organization international survey of adolescents that uses primarily standardized, self-administered questionnaires to make results comparable between countries. The core questionnaire looks at 10 domains of key factors affecting adolescent health, including tobacco use, alcohol abuse, drug use, diet, hygiene, physical activity, sexual behavior, violent behavior, and unintentional injuries, protective factors, and mental health. For the actual survey, the GSHS was approved by national governments and sponsored or organized by official agencies at the national level (usually by Ministry of Health or Education, and an institutional review board or ethical committee), using a school-class two-stage whole-group sampling method. Questionnaires were translated into the national language for student comprehension. After excluding the samples with missing data, countries covering the key variables of this study were selected, using the most recent data available for each country, and the final sample was drawn from survey data from 2003 to 2015, for a total of 167,286 samples from 65 countries, 5 regions (21,501 samples from Africa, 59,326 samples from Americas, 23,222 samples from Eastern Mediterranean, 13,301 samples from South East Asia, 49,936 samples from Western Pacific). --- Ethics Statement GSHS received ethics approval from each country. Written informed consent was obtained from participants or guardians prior to the survey, and privacy protections were obtained. The current study used publicly available data. --- Measures 2.3.1. Dependent Variable: Mental Health "Mental health": Mental health was measured based on the two indicators of loneliness and anxiety in the questionnaire [37], with the questions "During the past 12 months, how often have you felt lonely/been so worried about something that you could not sleep at night?". In order to visually explain the effect of bullying on adolescent mental health, this paper recoded the responses to the above measurement questions as "1 = always, 2 = most of the time, 3 = sometimes, 4 = rarely, 5 = never". The current methods for comprehensive index measurement include subjective weighting method and objective weighting method. The subjective weighting method determines the weight based on the researcher's subjective attention to the evaluation indicators, and the objective weighting method is based on the correlation between the indicators or indicators [38]. The degree of dispersion of information determines the weight. In order to eliminate the subjective arbitrariness of determining weights, following Huang's research [39], this study chose the entropy method in the objective weighting method to determine the weights between the various indicators of the observed variable "mental health", and used the information carried by the entropy value to calculate the "mental health". According to the entropy method, the study determined the weight of each index as follows: there were N samples and M evaluation indexes, which were the value of the j-th index of the i-th sample. Since the various indicators of sample mental health had different dimensions and orders of magnitude, the range standardized formula was used to process the indicators, and the absolute values of the indicators were converted into relative values to solve the homogeneity problem of various indicators, as shown below: Z ij = x ij -minx j maxx j -minx j The formula following shows the standardized value of the j-th index of the i-th sample, the minimum and maximum value of the j-th index. In order to ensure the non-negativity of the calculation results, this study would shift the coordinates by 1 unit: x ij = Z ij + 1 Finally, the standardized values were used to calculate the mental health "MH" of adolescents in the sample: MH = <unk> m i=1 w j P ij = <unk>[ 1 -e j <unk> m i=j 1 -e j <unk> x ij <unk> n i=j x ij ] The formula above represents the entropy value of the j-th index, k = 1/In(n); n is the sample size, which is the weight of the j-th index of the i-th sample. Finally, the evaluation score of "mental health" was calculated as 1.02-9.89, and the higher the score, the better the mental health status. --- Independent Variables: Frequency and Forms of Being Bullied "Frequency of being bullied": "Frequency of being bullied" was measured by the question "During the past 30 days, on how many days were you bullied?" and recoded (1 = 1 to 5 days, 2 = 6 to 19 days, 3 = more than 20 days). The larger value represented the higher frequency of being bullied. "Forms of being bullied": "Forms of being bullied" was measured by the question "During the past 30 days, how were you bullied most often?" and recoded (1 = physical bullying, 2 = verbal bullying, 3 = neglect). --- Control Variables Previous studies have reported that individual factors contributing to adolescent mental health, such as age, gender [40], substance use [26,41], weight status [42] and family socioeconomic status [43]. Therefore, the study used the following variables related to mental health of adolescents in GSHS as control variables, including age, gender, physical well-being, cigarette smoking, alcohol use, proxy of family socioeconomic status, number of close friends and frequency of missing school. "Weight status" was measured by the value of body mass index (BMI), calculated with two adolescents' indicators of height and weight, and recoded (1 = underweight, 2 = normal weight, 3 = overweight) [44,45]. "Cigarette smoking" and "alcohol use" were measured by the questions "During the past 30 days, on how many days did you smoke cigarettes / have at least drink containing alcohol?" and recoded (1 = less than 5 days, 2 = 6-19 days, 3= more than 20 days). According to a prior study [14], "proxy of family socioeconomic status" was measured by the question "During the past 30 days, how often did you go hungry because there was not enough food in your home?" and recoded (1 represents "low", 5 represents "high"). The larger value represented the higher family socioeconomic status. "Number of close friends" was measured by the question "How many close friends do you have?" (0 = 0 friends, 1 = 1 friend, 2 = 2 friends, 3 = 3 or more friends). "Frequency of missing school" was measured by the question "During the past 30 days, on how many days did you miss classes or school without permission?" (1 = less than 2 days, 2 = 3-9 days, 3 = more than 10 days). --- Protective Factors: Parental Supports Protective factors were assessed by parental supports. As critical factors of resiliency, parental supports included parental supervision, parental connectedness and parental bonding [34,35], based on the questions "how often did your parents or guardians check to see if your homework was done?", "how often did your parents or guardians understand your problems and worries?", and "how often did your parents or guardians really know what you were doing with your free time?", and assessed by frequency in the past 30 days (1 represents "never", 5 represents "always"). --- Statistical Analysis Firstly, the study conducted descriptive statistics on the overall prevalence of maltreatment and the prevalence of different forms of maltreatment among adolescents aged 12-17 years in 65 sample countries among five regions, and to visualize the differences in the distribution of bullying across regions, a global distribution of bullying rates among adolescents in the 65 sample countries was drawn. Secondly, an ordinary least squares model was used to analyze the effects of bullying frequency and different forms of bullying on adolescent mental health. In the model estimation, state-fixed effects were controlled for in addition to the effects of the above-mentioned control variables on adolescent mental health. The study further regressed subgroups by gender and age to estimate differences in the effects of bullying exposure, bullying frequency, and forms of bullying on adolescent mental health by gender and by age (under 15, over 15) across continents, respectively. The reason for choosing 15 years as the age group cut-off was that in most countries, adolescents under 15 years are at the middle school level and those over 15 years are at the high school level, where they show more significant differences in psychological and behavioral aspects [28]. Finally, the study conducted moderation test to explore the protective factors of adolescent bullying, i.e., whether parental support could play a significant positive role in the effect of bullying on adolescent mental health. The study used Stata 15.0 to analyze the data and ArcGIS software for mapping. --- Results --- Sample Description The descriptive statistics of the sample are shown in Table 1. The mean age of the sample adolescents was 14.14 years (SD = 1.20), of which 46.74% were male (78,187) and 53.26% were female (89,099). In terms of bullying prevalence, 32.03% of the 167,286 overall samples of adolescents aged 12-17 years had experienced bullying in the past 30 days of the survey. Regarding the frequency of bullying, 24.68% of adolescents were bullied for less than five days, less than 10% of adolescents were bullied for more than five days. In terms of mental health, the mean of mental health of adolescents in the sample countries was 5.79 (SD = 1.82), which was in the middle to upper level. Among different regions, the mental health level of adolescents in South East Asia was the highest (M = 5.97, SD = 1.78), and African adolescents' mental health level was the lowest (M = 5.47, SD = 1.92). In terms of parental support, the mean values of "parental supervision", "parental connectedness", and "parental bonding" for the overall sample of adolescents were 2.94 (SD = 1.49), 3.00 (SD = 1.46), and 3.19 (SD = 1.44), respectively. The mean values of "parental supervision" ranged from "rarely" to "sometimes", and the mean values of "parental connectedness" and "parental bonding" ranged from "sometimes" to "most of the time". --- Prevalence and Forms of Bullying across the Regions Table 1 shows the prevalence of different forms of bullying in the six regions. Overall, verbal bullying had the highest prevalence (66.36%), followed by physical bullying (24.02%), and the neglect had the lowest prevalence (9.62%). Across regions, physical bullying was highest in Africa (28.98%) and lowest in the Americas (18.84%); verbal bullying was the opposite of physical bullying, highest in the Americas (71.09%) and lowest in Africa (61.75%); neglect was highest in South East Asia (11.10%) and lowest in Eastern Mediterranean (7.11%). Table 2 and Figure 1 show specifically the prevalence of bullying and different forms of bullying in each sample country. In terms of bullying prevalence, the African region had the highest prevalence of adolescent bullying at 47.36%, followed by Eastern Mediterranean (41.53%), South East Asia (33.19%), Western Pacific (27.58%), and the Americas (26.23%). In terms of sample countries, 5 of the 12 sample countries in Africa had more than half of the bullying prevalence, namely, Botswana (52.20%), Ghana (56.72%), Kenya (54.35%), Zambia (61.58%), and Zimbabwe (59.15%). In Americas, the prevalence of bullying was ranging from 47.14% in Peru to 12.50% in Barbados. The Eastern Mediterranean region had the highest bullying rate in the Occupied Palestinian Territory with over half (52.27%) and the lowest bullying rate in Iraq (27.45%). South East Asia also had more than half of adolescents in Indonesia (50.14%) as its highest bullying rate, and the lowest adolescent bullying rate was in Myanmar (19.51%). Samoa, in the Western Pacific region, had the highest prevalence of bullying among all countries in the sample, at 72.40%, while Malaysia (16.99%) had the lowest prevalence of bullying among adolescents in the Western Pacific region. --- Effects of Bullying on Psychological Well-Being of Adolescents After controlling for state-fixed effects, the study used OLS models to examine the effect of bullying and different forms of bullying on adolescent mental health, and the results are shown in Table 3. In Model 1, with bullying frequency as the core explanatory variable, the regression results showed that bullying frequency negatively affected ado-lescent mental health, with the largest negative effect on mental health for adolescents who had been bullied for more than 20 days in the past 30 days, with a 7.53 decrease in mental health (p <unk> 0.001, CI: -7.72, -7.33). Model 2 further estimated the effects of different forms of bullying on adolescent mental health, and the results showed that verbal bullying negatively affected adolescent mental health mostly, with a 9.64 decrease in mental health (p <unk> 0.001, CI: -9.89, -1.01). Physical and neglect also negatively affected adolescent mental health, with a 7.49 (p <unk> 0.001, CI: -7.89, -7.10) and a 1.21 (p <unk> 0.001, CI: -1.27, -1.15) decrease in mental health, respectively, which verified H1. --- Effects of Bullying on Psychological Well-Being of Adolescents After controlling for state-fixed effects, the study used OLS models to examine the effect of bullying and different forms of bullying on adolescent mental health, and the results are shown in Table 3. In Model 1, with bullying frequency as the core explanatory variable, the regression results showed that bullying frequency negatively affected adolescent mental health, with the largest negative effect on mental health for adolescents who had been bullied for more than 20 days in the past 30 days, with a 7.53 decrease in mental health (p <unk> 0.001, CI: -7.72, -7.33). Model 2 further estimated the effects of different forms of bullying on adolescent mental health, and the results showed that verbal bullying negatively affected adolescent mental health mostly, with a 9.64 decrease in mental health (p <unk> 0.001, CI: -9.89, -1.01). Physical and neglect also negatively affected adolescent mental health, with a 7.49 (p <unk> 0.001, CI: -7.89, -7.10) and a 1.21 (p <unk> 0.001, CI: -1.27, -1.15) decrease in mental health, respectively, which verified H1. --- Effects of Bullying on Psychological Well-Being in Adolescents of Different Gender Table 5 demonstrates the effects of bullying frequency and bullying form on the mental health of adolescents by gender. Overall, both bullying frequency and bullying form had a significant negative effect on both male and female adolescents in the sample across continents (p <unk> 0.001). In the total sample, the negative effect of bullying frequency on the mental health of female adolescents was more significant than that of males (p <unk> 0.001). Specifically, the negative effect of bullying frequency on the mental health of female adolescents was greater than that of males in the sample countries of the Eastern Mediterranean region, the South East Asian region, and the Western Pacific region; the negative effect of bullying frequency on males was greater when the bullying frequency was less than 19 days in the sample countries of the American region (p <unk> 0.001). Looking at the different forms of bullying, verbal bullying and neglect had a greater negative effect on overall female adolescents than on males, while physical bullying had a greater negative effect on overall male adolescents, supporting partial of H2. Across continents, all three forms of bullying had a significant negative effect on the mental health of male adolescents in Africa compared with females (p <unk> 0.001); in the Americas, physical bullying had a greater negative effect on the mental health of male adolescents than females (p <unk> 0.001), neglect had a greater negative effect on the mental health of female adolescents than males, and verbal bullying did not differ between the two sexes; in the Eastern Mediterranean region, physical bullying and verbal bullying had a greater negative effect on females than males (p <unk> 0.001), and neglect had a more severe negative effect on males; in South East Asia, both physical bullying and verbal bullying had a more severe negative effect on females (p <unk> 0.001), and neglect had a more severe negative effect on males (p <unk> 0.001); in the Western Pacific, physical bullying and neglect had a more severe negative effect on female mental health, and verbal bullying had a more severe negative effect on male mental health. --- Effects of Bullying on Psychological Well-Being in Adolescents of Different Ages Table?? demonstrates the effects of bullying and different forms of bullying on the mental health of adolescents of different ages across continents and their variability. In terms of bullying frequency, bullying frequency had a greater negative effect on the overall mental health of adolescents under the age of 15 than adolescents over the age of 15 (p <unk> 0.001). In terms of forms of bullying, physical bullying, verbal bullying and neglect had a greater negative effect on the mental health of adolescents under 15 years old than adolescents over 15 years old in the total sample (p <unk> 0.001) as hypothesized. Among the regions, the negative effect of neglect on the mental health of adolescents over the age of 15 was more significant in the sample countries of the Western Pacific region (p <unk> 0.001), and the negative effect of physical bullying and verbal bullying on the mental health of adolescents under the age of 15 was more significant (p <unk> 0.001); the situation in the other regions was consistent with that of the overall sample. --- The Protective Effect of Parental Support on the Psychological Well-Being To test the potential moderating role of parental support as a protective factor on adolescent mental health after bullying, the study conducted the procedure to test significant interactions. The results from Table 6 show that being bullied was negatively associated with mental health in three models (p <unk> 0.001). Significant interaction effects between parental supervision and being bullied (p <unk> 0.001), between parental connectedness and being bullied (p <unk> 0.001), between parental bonding and being bullied (p <unk> 0.001) were found to be positively associated with psychological well-being, indicating that the moderating effect of parental support occurred in the protection of mental health of adolescents who experienced being bullied as H3 hypothesized. Table 7 shows the results of the effect of parental support on the mental health of adolescents following different forms of bullying. Among them, "parental connectedness" had a positive protective effect on the mental health of adolescents after verbal bullying or peer neglect, i.e., the more parents understand the adolescents' distress after verbal bullying or neglect at school, the higher the level of mental health of the adolescents, and the frequency of parental understanding increases by one unit, the level of mental health increased by 8.71 units (p <unk> 0.001) and 1.05 units (p <unk> 0.001), respectively; "parental bonding" had a positive restorative effect on the psychological health of adolescents who were verbally bullied, i.e., for each unit increase in the frequency of "parental bonding", the psychological health level of adolescents who were verbally bullied increased by 2.47 units (p <unk> 0.05). --- Discussion The study examined the overall prevalence of bullying among adolescents and the prevalence of different forms of bullying in a total of 167,286 sample in five regions, and further analyzed the effect of different forms of bullying on adolescent mental health, the protective role of parental support, and the main findings were as follows: Firstly, adolescent bullying cannot be ignored, with the highest prevalence of verbal bullying. Our study showed that the overall prevalence of bullying among adoles-cents in the 167,286 sample countries was 32.03%, a result that was consistent with the previous UNICEF report published in 2018 that more than one-third of students aged 13-15 worldwide experienced bullying. The results of Biswas et al. (2020) and Elgar et al. (2015) cross-regional comparative studies on bullying and violence among adolescents were generally consistent with the results of the two studies on the prevalence of bullying among adolescents, which were 31% [32] and 30% [8], respectively. From the results of the cross-regional comparison, the highest prevalence of bullying among adolescents (47.36%) was found in the sample countries in the African region, which may be related to the low-income level, poorer schools, and social environment, war, and riots in the African region [46]. In terms of the prevalence of different forms of bullying, verbal bullying had the highest prevalence (66.36%), followed by physical bullying (24.02%), and neglect had the lowest prevalence (9.62%). The results of a survey conducted by Scheithauer et al. (2006) in Germany with students in grades 5-10 [47], and the results of the prevalence of six forms of bullying among 2667 Italian secondary school students, obtained by Vieno et al. in 2011 using the results of the Health Behavior in School-aged Children Survey database, also both showed the highest prevalence of verbal bullying, consistent with the findings of this paper [48]. This suggested that verbal bullying, which takes the form of making fun of a peer's race, nationality, color, creed, body, and appearance, was the most prevalent and most likely to occur among adolescents because it was the most recognizable and less costly to occur. However, it was worth pointing out that the findings for the prevalence of physical bullying and neglect in this study differ slightly from those of the two studies mentioned above, due to the different criteria used to measure them. Secondly, compared with physical bullying and neglect, verbal bullying had the most serious negative effect on adolescent mental health. Not only did verbal bullying had the highest prevalence of the three forms of bullying, but it also had the most serious negative effect on adolescent mental health for two main reasons: firstly, verbal bullying occurred most frequently, and according to the study, the frequency of bullying significantly and negatively affects adolescent mental health, so the lower the level of mental health when adolescents suffered frequent ridicule or name-calling from peers; secondly, from the perspective of social identity theory, this highly discriminatory ridicule led to negative mental health outcomes, especially for adolescents with extremely strong identity, and this discrimination increased their psychological distress [49]. Thirdly, overall, the frequency of bullying had a greater negative effect on the mental health of female adolescents compared with male adolescents, which was consistent with the findings of a recent study conducted in the United States that school bullying had a greater effect on psychological depression in females than in males [26]. In addition, physical bullying had a greater negative effect on the mental health of male adolescents, and verbal bullying and neglect had a greater negative effect on the mental health of female adolescents. This was generally consistent with previous research finding that depressive symptoms were more pronounced after active forms of bullying (i.e., physical bullying) in boys and after passive forms of bullying (i.e., verbal and relational bullying) in girls [28,50]. This would require further exploration of the effect of different forms of bullying on the mental health of male and female adolescents in specific regions. While we need to protect boys and girls equally from bullying, countries also need to consider the gender differences in the occurrence and effect of different forms of bullying in their countries and pay targeted attention to adolescents who are victims of bullying. Fourth, the frequency of bullying had a more significant negative effect on the mental health of adolescents under the age of 15, and different forms of bullying also had a more significant negative effect on the mental health of adolescents under the age of 15. Previous studies have found that the odds of bullying are higher for younger adolescents (under 15) [25,51]. Compared with younger adolescents, older adolescents (15 years and older) were more aware of self-concept and self-regulation in terms of self-perception and psychological construction [49], so both the frequency of bullying and the different forms of bullying had a more significant negative effect on the mental health of adolescents under the age of 15. In addition, the study showed regional differences in mental health of adolescents in different age groups after various forms of bullying, which provided a basis for the development and implementation of intervention policies in each region or country. Finally, in terms of protective factors, "parental supervision", "parental connectedness" and "parental bonding" played positive roles in the relationship between bullying and adolescent mental health. Positive relationships, especially positive family relationships that provided intimacy, support, trust, emotional comfort, and a sense of belonging, are one of the key elements of resiliency [52]. In such a family environment, even if adolescents were abused and bullied, they could still buffer the stress and shock from other aspects by increasing their self-efficacy, self-worth, and emotional belongingness [53]. "Parental connectedness" and "parental bonding" were important indicators of parent-child intimacy and emotional comfort, and played a positive role in adolescents' resilience. However, there were no consistent conclusions to the role of "parental supervision". Some studies have not found a significant link between parental supervision and mental health after bullying [54,55]. Others have identified the lack of parental supervision as a risk factor to adolescents' mental health development [56], which is consistent with the current study. Future research would explore how the degree or the forms of parental supervision influence mental health when adolescents experience bullying. Limited by the consistency of the GSHS database, this study suffered from the following shortcomings: Firstly, the countries or regions selected represent only some of the five regions. We did not contain the European continent because only one country provided useful data. Future studies would include more specific countries to explore the global adolescent bullying situation. Secondly, the GSHS used a self-administered questionnaire, and although self-administration was an acceptable way to collect data on adolescent bullying victimization, there was a limitation of possible shared method variance. Finally, we observed significant regional differences in the prevalence of different forms of bullying, including gender differences and age differences, and future research would consider social context and cultural heterogeneity to explain regional differences better and provide more possibilities for countries to implement adolescent bullying intervention programs. --- Conclusions Despite these limitations, our study contributed to the exploration of adolescent bullying in the following ways: firstly, unlike previous studies limited to individual countries or regions, our analysis covered 65 sample countries across five continents, providing more evidence for cross-regional comparative studies of adolescent bullying; secondly, in addition to focusing on bullying among adolescents as a whole and its effect on mental health, we focused on intergroup differences in adolescent subgroups (gender groups and age groups) to provide a basis for targeted development of specific intervention policies for different groups of adolescents. Finally, we focused on the potential protective factors of adolescent bullying and found that "parental supervision", "parental connectedness" and "parental bonding" played a positive role in protecting the psychological health of adolescents who were bullied. The above findings suggested that, as a global public health problem, adolescent bullying should attract sufficient policy concern and practical intervention, and further establish a comprehensive adolescent social protection mechanism and protection system including family, school, and community. --- Data Availability Statement: The data are available online at https://www.cdc.gov/gshs/ (accessed on 10 December 2021). --- Author Contributions: Conceptualization, X.M.; methodology, X.M. and J.L.; software, Z.X. and J.L.; validation, X.M., J.L. and Z.X.; formal analysis, X.M., J.L. and Z.X.; data curation, Z.X.; writing-original draft preparation, X.M. and J.L.; writing-review and editing, X.M., J.L. and Z.X.; visualization, X.M., J.L. and Z.X.; supervision, X.M.

Adolescent bullying is a public health issue of great global concern. Given the serious negative effect of bullying on adolescent mental health, it is critical to seek protective factors to protect adolescent mental health. From a global cross-regional perspective, the study aims to explore the relationship between forms of bullying and adolescent mental health and the role of parental support as a protective factor. Data were drawn from adolescents aged 12-17 years in 65 countries from the Global School-based Student Health Survey between 2003 and 2015. After controlling the state-fixed effects, individual adolescent behavior, and family factors, the ordinary least squares model was used to analyze the influence of bullying frequency and forms of bullying on adolescent mental health. The results found that the prevalence of bullying in the sample of 167,286 adolescents was 32.03%, with the highest prevalence of bullying in the sample countries in Africa. Verbal bullying had the highest prevalence and the most significant negative effect on adolescent mental health. The study also discussed the differences in bullying among adolescents by gender, age, and region. "Parental supervision", "parental connectedness" and "parental bonding" played a positive and protective role in the mental health of adolescents who experienced bullying.

in protecting the psychological health of adolescents who were bullied. The above findings suggested that, as a global public health problem, adolescent bullying should attract sufficient policy concern and practical intervention, and further establish a comprehensive adolescent social protection mechanism and protection system including family, school, and community. --- Data Availability Statement: The data are available online at https://www.cdc.gov/gshs/ (accessed on 10 December 2021). --- Author Contributions: Conceptualization, X.M.; methodology, X.M. and J.L.; software, Z.X. and J.L.; validation, X.M., J.L. and Z.X.; formal analysis, X.M., J.L. and Z.X.; data curation, Z.X.; writing-original draft preparation, X.M. and J.L.; writing-review and editing, X.M., J.L. and Z.X.; visualization, X.M., J.L. and Z.X.; supervision, X.M.; project administration, X.M. All authors have read and agreed to the published version of the manuscript. --- Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. --- Conflicts of Interest: The authors declare no conflict of interest.

Adolescent bullying is a public health issue of great global concern. Given the serious negative effect of bullying on adolescent mental health, it is critical to seek protective factors to protect adolescent mental health. From a global cross-regional perspective, the study aims to explore the relationship between forms of bullying and adolescent mental health and the role of parental support as a protective factor. Data were drawn from adolescents aged 12-17 years in 65 countries from the Global School-based Student Health Survey between 2003 and 2015. After controlling the state-fixed effects, individual adolescent behavior, and family factors, the ordinary least squares model was used to analyze the influence of bullying frequency and forms of bullying on adolescent mental health. The results found that the prevalence of bullying in the sample of 167,286 adolescents was 32.03%, with the highest prevalence of bullying in the sample countries in Africa. Verbal bullying had the highest prevalence and the most significant negative effect on adolescent mental health. The study also discussed the differences in bullying among adolescents by gender, age, and region. "Parental supervision", "parental connectedness" and "parental bonding" played a positive and protective role in the mental health of adolescents who experienced bullying.

Introduction: Celebrating 175 Years of Organized Diaconia in Germany During the WCC 11th Assembly in Karlsruhe, Germany, in August/September 2022 a new 140-page study document was introduced to and discussed by the ecumenical fellowship of churches, published under the title "Called to Transformation-Ecumenical Diaconia" 1. This article introduces the context, major content, and key convictions of this study document and related discussions in Karlsruhe. It answers the question: what is the significance of the global ecumenical fellowship of churches in focusing on a common language of diaconia and a more explicit church-related terminology to describe and strengthen the social services of churches and their engagement in diaconia and social witness? German churches are celebrating a significant jubilee in 2023, related to 175 years of organized social work of Protestant churches in the country. This is also a special occasion for contemplating the relevance of the strong connectivity between churches and diaconia from a global perspective. A campaign titled "Out of Love: Diaconia" 2 was created, and resources are being made available to commemorate the impact of the famous Wittenberg 1848 speech of Johann Hinrich Wichern. During the national convention of Protestant churches in Germany in that year, he criticized churches for their failure to address the needs and concerns of the impoverished urban masses resulting from rapid industrialization and the Napoleon wars in the first half of the 19th century. 3 His strategic recommendation to form a national platform and joint committee of organized social care ("Central Ausschuss für Innere Mission") laid the basis for strong connectivity between churches and diaconia in Germany for the subsequent decades. Decades later, this led to the formation of the Protestant agency for Diaconia and Development (Evangelisches Werk für Diakonie und Entwicklung, EWDE), which has had its headquarters in Berlin since 2012; with its 800 employees, it can be regarded as the biggest Christian NGO in Europe. Protestant diaconia in Germany in the 175 years of its history has become a brand name and a strong agency alongside and within local churches. With 600,000 employed workers and 700,000 volunteers, it is one of the largest and widespread employers in German society, today operating within the stable framework of state-church partnership relationships and with a major part of its services offered for the poor, suffering, and needy, considerably co-funded by the state authorities within the German welfare system (Hübner 2021). --- Internationalizing Diaconia after the Second World War What is less known than the history of national diaconia within Germany is the parallel history of international or worldwide diaconia that became called "ecumenical diaconia" in the dialogue with the global ecumenical movement (Werner and Ross 2021). Whereas diaconia (or "Christian love of the neighbor" as it was called in the earlier stages) first and foremost was always a response of Christian communities to situations of need in the immediate context of local environments, it is clear already in apostolic times that there was also a realization of the churches' responsibility to respond to situations of need in the far distance (for example, collect for Jerusalem ((1 Cor 16:1-4; 2 Cor 8:1-9:15; Rom 15:14-32); famine in Judea (Acts 11,27-29)). Christianity, from its founding periods, remained aware that God's promise and rule of mercy is the whole earth (Ps. 24,1), the global inhabited space (the oikumene), and that all of those in need amongst the "least" ask for our response (Mat. 25,40). With interchurch aid having long been a structural component of diaconia, the main unleashing factor for the discovery and unfolding of ecumenical diaconia as intentional assistance to people in need and populations afar and even beyond membership in the church was the ecumenical movement in the early part of the 20th century. The last decades of the 19th century and the beginning of the 20th century saw a growing concern for Christian social thinking and the first pioneering conferences for ecumenical social ethics, particularly the Stockholm conference in 1925, which will be commemorated by the Church of Sweden and its ecumenical partners in 2025. 4 The encounter both with massive destruction and refugee movements in the aftermath of the two World Wars and the subsequent widening of horizons of the ecumenical movement to include issues of unequal distribution of resources in the period of decolonialization paved the way to a strong commitment to the ecumenical and international dimension of diaconia. In 1945, all Protestant churches in Germany came together to found the Office of Inter-Church Aid (Hilfswerk der Evangelischen Kirchen in Deutschland) (Füllkrug-Weitzel 2016). In the same year (three years prior to the formation of the WCC), the Division on Inter-Church Aid, Refugee and World Service (later called Commission on Inter-Church Aid, Refugee and World Service (CICARWS)) was established in Geneva as part of the "WCC in formation". Both can be understood only in the context of the huge refugee crisis and the urgent need for the reconstruction of European countries at that historical moment. Germany, at that stage, had to cope not with 1 million refugees (as in 2015), but with 15 million refugees. Thus, the early ecumenical movement in the post Second World War period was confronted with two major challenges, for which two distinct early instruments of ecumenical diaconia were developed (although the terminology was still developing in that period): The first major challenge was the reconstruction of Europe and the refugee crisis: CICARWS was the main instrument of engaging in interchurch aid and ecumenical diaconia in this regard. The second major challenge that soon followed was the one related to the processes of decolonialization and the growing gap between rich and poor countries in the Global South. The main instrument of response developed in this regard was the missionary assistance program of the International Missionary Council (IMC), which has existed since 1922. A major paradigm shift emerged with the formal integration of the IMC into the WCC in 1961 during the assembly in New Delhi: now the ecumenical movement became truly global. The work in interchurch aid irreversibly was enlarged by newly emerging issues concerning development, liberation, and structural economic injustice. It is interesting to see that with the integration of the International Missionary Council (IMC) and the WCC, the mandate of CICARWS was also changed to combine interchurch aid with global diaconia or ecumenical diaconia: in 1961, the new mandate of CICARWS was stated as follows: The aim of the division shall be to express the ecumenical solidarity of the churches through mutual aid in order to strengthen them in their life and mission and especially in their service to the world around them (diaconia) and to provide facilities by which the churches may serve men and women in acute human need everywhere, especially orphaned peoples, including refugees of all categories. (See Murray 1961, p. 7; also quoted in Robra 1994) The growing awareness of the mutual relationship between ecumenism and diaconia in the 1950s and 1960s stimulated a process of a gradual widening and merger of both agendas (interchurch aid in Europe and missionary assistance in the global context) and their related instruments. In 1957, an international consultation held in the Evangelical Academy of Berlin introduced the concept of ecumenical diaconia and deepened the conceptualization of this new understanding. It reflected a strong sense of postwar Christianity in its commitment to struggle for global peace and overcoming hunger on global scale and to hold together the vision of a'responsible society' and the churches' commitment to'social diaconia', both to be understood in a global perspective, hence the new emphasis on ecumenical diaconia (Akademie 1957; Berg 1959). In 1959, a national collection was organized by all Protestant churches in Germany to provide support for the global poor, mainly in west Africa at that time (Biafra crisis). Bread for the World, the Protestant development agency in Germany, was born out of this globalized concern for the poor and a passion for new shapes of ecumenical diaconia (Füllkrug-Weitzel 2016). Having received ecumenical aid and support from churches in the west and south, Protestant churches in Germany wanted to become actors and drivers for solidarity in ecumenical diaconia on a permanent basis. The later rights-based approach to ecumenical diaconia then emerged as a direct consequence of the learning process in ecumenical social ethics to understand the structural root causes of rising inequality and global poverty. The commitment for a partner-centered working methodology, i.e., the need to collaborate and to support partner organizations from both church-based background and the secular agencies of civil society, which can render an effective contribution to the eradication of poverty and the improvement in human rights situations of women, children, and the marginalized, can be seen as a direct implication of the interest to support approaches in social development from below and to put people's interests first. --- Ecumenical Diaconia in the Context of the Discourse on the Development Paradigm in the Period of 1968 to 1991 Bread for the World grew in close collaboration with the global ecumenical instruments, which were created in the 1960s and 1970s in the spirit of a new sense of the global dimensions of a "responsible society": the IMC-based global study project Rapid Social Change (1955-1961) provided a first entry and opening of the growing debate of development issues. The 4th Assembly of the WCC (held in Uppsala, 1968) can be regarded as the beginning of a formal engagement of the institutionalized ecumenical movement in the global development discourse. The transformation of interchurch aid into development aid was critically discussed at that time, indicating that despite the changing political frame-work and new opportunities for cooperation between church-based ecumenical diaconia and state development programs (and funding), a theological concern remained vital in many circles to keep the churches' international work for social justice with a clear Christian profile, close to its original foundations rooted in the mandates for ecumenical diakonia, and not to lose its distinctive marks as a church based activity-a debate taken up today again with the new study document of the WCC and ACT Alliance on ecumenical diaconia after 2014. The Commission of Churches for Participation in Development (CCPD, founded in 1970) became a major partner for Bread for the World as well as the unfolding instruments of CICARWS as the project division for ecumenical diaconia in the WCC. The failure of the first development decade and its concept of development as modernization to acknowledge the relevance of the factor of religion in social transformation processes (Werner 2019; see also Tomalin 2015) and the growing disparities in terms of social progress, access to basic means of life, and the pervasive influence of postcolonial dependency, however, also provoked very critical questions with regard to the condition in which ecumenical diaconia had to take shape. Was a fair and mutual approach to ecumenical diaconia possible at all in a context that suffers from continued dependency between churches mainly in a giving role and other churches mainly in a receiving role? Who sets which priorities and formulates the conditions determining who sits at the table and in which role? This was the background for the study process on Ecumenical Sharing of Resources (ESR), which was initiated in the WCC after the 5th Assembly (held in Nairobi, 1975). 5 This was a reaction to the so-called "moratorium debate" on the sending of personnel and funds, raising critical questions about selfhood and the identity of churches in the Global South challenging the static division between and portraying of'receiving' churches and'sending' or 'giving' churches. The ESR study project aimed at a new framework of relationships that would free the churches from traditional roles of being either a sending (giving) or a receiving body and enable them to overcome structures of inequality and dependency between rich and poor. A broader understanding of what is meant by "resources" was projected, including spiritual, cultural, and human resources, as well as financial and material goods. The ESR study called for just relationships based on equality, which would allow for mutual accountability, sharing of power, and true interdependence. These require a holding together of mission, development, and service, which were seen as sometimes moving apart and being treated separately, both in theology and organizational structures. The study materials of the ESR process (WCC 1980) are still worth rereading as they include questions that are still relevant. However, even the guidelines on ecumenical resource sharing finally presented at the WCC (WCC 1988) could not present a breakthrough in the mechanisms of ecumenical diaconia on a broader and international scale, as too many partners were involved and much of the implementation of these goals took place in regional or bilateral networks or international mission boards (such as the United Evangelical Mission (UEM) 6 ; Communeauté d'Eglise en Mission (CEVAA) 7, and others). In the same time period, the Nairobi Assembly in 1975 replaced the term "development" and deliberately articulated a new umbrella concept of a just, participatory, and sustainable society (JPSS) to inspire ecumenical diakonia, which presented a definite break with the equation of "development" and economic "growth and modernization" (see also ACT Alliance 2008) and was a response to the growing recognition of the persistence of poverty and misery and of the limits of and threats to the Earth's capacity to sustain human life. --- Bringing Churches and Agencies Together Again: The Study Process on Ecumenical Diaconia after 2014 The growth and further professionalization of church-related development agencies and humanitarian actors in the 1990s led to the formation of a global network of agencies for ecumenical diaconia, called ACT Alliance or Action by Churches Together for Development. It is the result of a merger in 2010 of ACT International, which was formed in 1995 to work on delivering humanitarian aid, and of ACT Development, which was formed in 2007. ACT Alliance 8 now presents a strong institutional presence of specialized ministries at the international level, which works in close collaboration with the WCC 9 but no longer as one of its subunits (as had been the case with CICARWS): around 130 specialized ministries and churches are now joined together in this international umbrella organization, which provides strong visibility of ecumenical diaconia in the global world of development and humanitarian actors at the UN level as well as in several regions. While it could be argued that the institutionalization of ACT Alliance presented, to a certain degree, a separate institutionalization of ecumenical diaconia over the ecumenical fellowship of churches (and their primary instruments of national or regional councils of churches), which was seen as necessary due to the demands of effectiveness, rapid reaction, and decision-making mechanisms particularly in international humanitarian work, a countermovement could be sensed soon in the debates, which started in the second decade of the new century. In the international dialogue in the second decade of the 21st century, one could sense increasing signals of a renewed interest in the ecumenical rediscovery of the ancient tradition of diaconal witness and services, particularly in churches in the nonwestern world (Ham 2015; Dietrich et al. 2015). Since 2012, international gatherings and consultations have called for a new attention and greater emphasis on the diaconal ministry of the church and the relevance of diaconia and advocacy in theological training and ministerial formation. 10 There is a long tradition of reflecting on diaconia and diaconate in the history of the World Council of Churches (White 2002). In the ecumenical missiological dialogue, the renewal of diaconia had also become a major issue, focusing on diaconia from the perspectives of the marginalized (WCC 2012). Some church traditions since then have developed explicit framework documents to describe theological and biblical perspectives on their understanding of diaconia 11, whereas in other traditions, the term and concept of diaconia (some use also the language of "social services" or social ministries) is used but does not yet have so much common conceptual grounding (Lutheran World Federation (LWF) (2009); Robra 2005). However, in many churches, both in the nonwestern world as well as in some western countries, training capacities, theological resources, and deliberate planning for diaconal capacities both in local churches as well as in diaconal agencies are still at low levels or given very unequal priorities. Conversely, there was a huge interest in learning and rediscovering the nature and ancient traditions of diaconal service in churches in multireligious contexts, both in Asian and African contexts. Kjell Nordstokke argued that in much of the ecumenical discourse so far, the reflection on diaconia and diaconate has followed two separate tracks: the first considering diaconia as interchurch aid in contexts of human need, while the entry point of the second has been the ecumenical perception of the church's ministry and, within it, a possible renewal of the diaconate. Nordstokke argued that these two tracks, to a large degree, have not been interrelated but have mainly remained isolated from each other. This has limited the reflection on both concepts; consequently, both had lost momentum in ecumenical theology (Nordstokke 2011, 2013). As a consequence of a longer process of discussing relationships both of the estrangement as well as of a new rapprochement and interaction between churches, specialized ministries and ecumenical organizations (the WCC, ACT Alliance, and LWF) had planned a major new learning process on key issues of ecumenical diaconia in Malawi in September 2014. 12 This international consultation emphasized that a strong relationship needs to be built between churches and specialized ministries in order to express a visible commitment of churches to peace, justice, and sustainability in societies. Specialized ministries were called to further develop ecumenical solidarity and to strengthen cooperation with churches. It is a result of this process that, some years later, the first version of a major study document was published both by the WCC and ACT Alliance on ecumenical diakonia, which also served as a background document for a back-to-back conference of the ACT Assembly and the WCC Executive Meeting in autumn 2018 in Sweden (WCC 2017). The basic intention of this study document was to understand diaconia as a dimension integral to the nature and mission of the church as a whole at all levels of its existence. Diaconia was viewed as a key expression of the strong link between what the churches are and what they do. Reflection on ecumenical diaconia requires understanding of both dimensions: the churches' being and their joint action as a worldwide communion of Christians and institutions. 13 After a special working commission on ecumenical diaconia from the WCC and ACT Alliance reworked this first version of the study document in order to deal with some criticism of orthodox member churches of the WCC (Phiri 2018) and after adding two substantial chapters on ecumenical diaconia in confessional contexts (chapter 8) and ecumenical diaconia in regional contexts (chapter 9), the 140-pagesstudy document was released in 2020 and made officially available to the 11th assembly of the WCC meeting in Karlsruhe in August/September 2022 under the title Called To Transformation-Ecumenical Diaconia(WCC-ED) (WCC 2022). 14 It is accompanied by major reflections and an addendum on the diaconal work of churches in the context of COVID-19 (WCC 2022, pp. 116-36). The following ten short reflections (paragraphs no. 5-15 in the subsequent sections) shed some light on the characteristic key convictions and main reflections from this historic paper, which is the first comprehensive and foundational theological study document on diaconia for the level of world Christianity. --- The Potential of the New, Yet Ancient Semantics of Ecumenical Diaconia During the New Delhi assembly of the WCC in 1961, there were only three major thematic sections planned for the assembly, very simply identified as Unity, Service, and Mission. This convincing thematic structuring of the assembly was not just for technical and administrative facilitation or reasons of brevity: it was a foundational theological statement at that early period of the global ecumenical movement expressing the fundamental conviction: the three original streams of the global ecumenical moment, namely faith and order, life and work, and world mission inseparably belong together and complement each other in their integrity. There is no ecumenical movement without the service dimension, i.e., the dimension of diaconia. While diaconia serves as a publicly known brand name only in some western European as well as Scandinavian churches, while other churches use different terminologies to refer to the phenomenon of social presence of Christian faith (such as integral mission, social service, and philanthropy), there are common ancient roots of all Christian churches that relate their social witness to a semantic family of words rooted in some derivates of diakonein. The cloud of words belonging to the diak-root-meaning serving others, freeing slaves, applying social protection laws for the needy, and establishing regular support systems for widows, orphans, women, and the elderly-according to the WCC-ED study document, belong to the very essence of what it means to be the church and to act as a church. As much as the missionary success and credibility of the ancient church was very much due to its diaconal inclusivity, the diaconal language was indispensable and remains of pivotal importance in today's secularized or multifaith environments for the credibility of the churches' witness and service. The three classical forms of interchurch aid, regular social support for the marginalized, as well as humanitarian action in emergency situations showed up as indispensable dimensions of being the church already in ancient biblical tradition (collect for Jerusalem and emergency aid during famine as cited before). The return to this precious common biblical language of church involvement in social services, action, and advocacy sends an important signal of a strengthened ecclesial identity of social action and critical Christian participation in development: While the churches are certainly and actively joining modern development discourses in debates about the understanding of the SDGs and the advocacy activities of civil society, their work cannot be completely and indiscriminately absorbed by these secular discourses. Instead, the social and political involvement of churches has its own specific language and profile by which it is linked together both horizontally and vertically to the global fellowship of churches, to the ecumenical dimensions of church history, and to the different current contexts of global Christianity. The ancient narratives of diaconal service, which are uniquely interwoven with the narratives of preaching and proclamation as the other ministry dimension of the church (see: book of Acts), remain significant for the identity and relevance of the global church. This is because these ancient diaconal narratives have communicative significance and relevance in any social and political context, be this rural or urban, modern or premodern, pre-or post-enlightenment, secular or nonsecular. Thus, the critical reminder of the common and ancient biblical language of diaconia is crucial and relevant for a process that is about the rediscovery and deepened understanding of the social relevance of the Gospel in contexts of nonwestern Christianity, particularly in those contexts where, for historical, political, or doctrinal reasons, it was not yet possible to develop large-scale diaconal institutions or even any social activity of Christians beyond the realm of private households. --- Realizing the Distinctiveness of Churches as Actors of Social-Ecological Transformation The joint WCC/ACT Alliance Document (WCC 2022) devotes a whole section to the issue of the Distinctiveness of Diaconal Action (WCC 2022, chp. 6). The interest of the WCC-ED document is that the churches remain distinct in their profile and that their added value as communities of faith becomes apparent. Church-related social services should be more than "Oxfam with church hymns", to quote a well-known slogan from the working group on ecumenical diaconia. The document is quite clear that this distinctiveness has to be repeatedly realized, and it is not sufficient to just claim this on a doctrinal or rhetorical level: To clarify the distinctiveness of diaconal action does not imply a claim of superiority, that diaconal agents are better and more trustworthy than other social agents. It simply means recognizing what has motivated Christians to take initiatives and establish institutions and organizations, and asking how their traditions and values, their bonds of social belonging and partnership, can be activated and become valuable assets in the daily performance of diaconal service. (see WCC 2022, chp. 6.1. ) Three distinct features on what churches can offer as a distinct profile and potential to bring into the processes of social and ecological transformation are named and spelled out here: (a) Churches are communities of hope, lived hope. Therefore, they are sleeping giants of social transformation that present additional resources for resilience of hope, which are related to their specific language, the language of liturgy and spirituality. This might not be easily measurable in terms of project outcomes and impact assessment, which we have become used to so much. Still, this spiritual resilience is essential as it provides us with a sense of dignity, of purpose, and of not serving in vain in contexts of suffering and desperation. The resilience of hope also needs to manifest itself, to take physical form, in'resilient communities'. Hope motivates and mobilizes communities to jointly come through times of adversary. This has relevance for all social transformation processes but also in the immediate survival phase of humanitarian disasters. Being a community of lived hope implies that involvement in development work or ecumenical diaconia can never just be reduced to project reporting, administrative monitoring, and impact-measuring paperwork, although these are relevant ingredients in successful and result-oriented development cooperation. These parameters are not everything and are insufficient alone. Otherwise, church-related development actors would become like any other department within a ministry of development cooperation or any secular organization of international humanitarian aid. Communication between churches and church-related faith actors needs time beyond project business, time for spiritual listening, for the fellowship of mutual praying and sharing one vision of hope together, for weeping together on the sorrows we experience, and enjoying together the successes with which we are graced. The quality time and spiritual depth of dialogical partnership work are some of the benefits underlining the distinct profile of churches' contributions to development work. The return to the biblical language of diaconia is a safeguard against the inherent danger to lose the ecclesial and prophetic dimensions of Christian social service against the secular-and often economic-growth-driven concerns for development. "The language of diaconia and of development represent different universes of discourse and practise which cannot be easily be brought together. Diaconia refers to a constitutive dimension of the life and mission of the churches as communities which are called to live as an active witness to the values of the kingdom of God. Development is a socio-political concept and economic concept which is rooted in the postcolonial efforts to improve the economic and social conditions in the newly independent countries in the southern hemisphere". (Raiser 2017, p. 2) (b) Secondly, churches are communities of Kingdom values, of spiritual resources that contribute to value change on wider social masses levels, which is so urgently needed for a new mentality of living within planetary boundaries. Nothing is more important today than to change the mentality of people, sharing in the common conviction that all have to change their lifestyles, their consumption patterns, their travel, mobility and traffic customs, their nutrition, and their production and waste management cycles in order to live within the planetary boundaries. The churches are therefore, to use a semantic that was developed in government and the German Ministry of Development Cooperation (BMZ) discourses on religion and development, important and indispensable factors of the cultural and value-driven dimension of eco-social transformation. Churches are also key players to resist the growing trends of shrinking spaces for civil society in a context where the intersection between faith and human rights (Prove et al. 2022) is sacrificed by autocratic governments that are not ashamed to support the recession of global democracy and the erosion of civic spaces to an extend that is unprecedented. 15 (c) Finally, churches are not just trapped and confined within the predominant discourse patterns of state development cooperation or the SDG agenda (as an example, see the ACT Alliance document "Our Understanding of Development" 2008, which later gave rise to the debate on ecumenical diaconia during the Busan assembly 2013) (Ham 2013). Churches are empowered and called to ask critical questions beyond the current national economic paradigm, even beyond the current SDG development framework concept and its assumptions on growth and progress: Do we have a really convincing new mega-narrative to be shared with populations that makes the imperative attractive for all to seek for a new balance between reducing our resources' depletion from this Earth and the needs to correct the huge global economic inequalities that still mark this world? Already, the current version of the ED Document is challenging the SDG framework for not having sufficiently integrated the issues related to the United Nations Declaration on the Rights of Indigenous Peoples (UNDRIP), which was a key focus at the Karlsruhe Assembly. 16 Some years ago, the Pope Francis 'Laudato Si' encyclical was moving toward something like a new mega-narrative, as it was about sharing what we have, in contrast to the Agenda 2030, which still presupposes a 'growing the cake', so that there is something to be shared for all (Hickel et al. 2023). We are still in the search for a global new mega-narrative that can bring people, nations, and civilizations together in one sense of direction, which does not allow for polarization on narrow-minded nationalistic or partisan interests. --- Nurturing Bilingual Approaches in Social Services and Advocacy-Keeping Faith-Based and Rights-Based Languages Together The WCC/ACT Alliance study document also favors the development of "bilingualism" in the understanding and practice of diaconia and social action: "What is distinct for diaconal actors is that they refer to religious concepts, in addition to secular, when explaining their action and its objective. This is an expression of the interdisciplinary nature of diaconia; it is social action rooted in socio-political knowledge and analysis, and, at the same time, it is faith-based and rights-based action." (WCC-ED No 6.2.). By using this bilingual approach, the WCC is intending to avoid both extremes: the self-secularization of a church that, while trying to be most effective and impact-oriented in terms of its concrete social action, tends to forget about the specificity of its own language and tradition, as well as the other extreme: to avoid the danger of a church becoming self-isolated or overspiritualized in its religious language, either by trying to protect its own limited existence or by defending the small individual and personal realm of faith. The point of this concern for stressing the spiritual nature of the church and its identity, and therefore the bilingual nature of the churches' engagement in social and ecological affairs, is to communicate who we are and what we do as churches and not to keep something for ourselves, which essentially belongs to ecclesial identity (mirroring Phil 2, 6-11... ). This is not opposed to churches opting for strong cooperation with secular development and social change actors (governments as well as intergovernmental organizations (INGOs) and the private sector). It remains important for churches to cooperate with nonreligious and other religious actors not only for effectiveness of their social witness but also for the secular actors to understand and appreciate the potential of local faith actors' engagement. This is why churches need to remain open and engaging with platforms such as Partnership on Religion and Development 17 (PaRD) and other secular actors in research and policy dialogue. Losing a distinct and proper church identity can thus take place in two different forms, both to be avoided, namely to overemphasize social relevance and assimilation to secular milieus as well as an overemphasis on self-protection and withdrawing from the political and social arenas. Emphasizing a bilingual approach and interdisciplinary competences in church-related international diaconia and humanitarian work nowadays means not to shy away from spiritual or religious language in cases and contexts where this is appropriate and supports the specific identity of church-based social action. The added value churches can bring to the social and political arena is because they speak a faith-based language, which then properly needs to be translated to ethical and political language. It certainly is known that that articulating faith language to the people can be quite challenging in religiously pluralistic contexts, as recipients of these social actions are often more interested in being supported than in their religious subscription and message of hope if they are not religiously inclined. However, the issues of explicit faith language have been quite vividly discussed in recent years, both with regard to the general strategy for humanitarian action and development 18 as well as, for instance, in the field of psycho-social care. 19 The specific value given here is that only faith-based language can engage with the fundamental layers of beliefs, values, and convictions, where social and political transformations must take root for the transformation to be lasting, i.e., sustainable. Faith language in long-term perspectives can thus be a decisive factor deciding about the sustainability of social-ecological transition processes. The language of ecumenical diaconia thus seriously considers the reality of more than 85% of the world's population, which, in its mindset, is coined and framed by different forms of spirituality: Most of the people in the world are religious. Diaconia must take this reality seriously. While Ecumenical diaconia does not intend to proselytize or work towards conversion, it rather recognizes the spiritual needs and context of the world, which is multi-religious. While doing interventions in the world, spiritual needs of the people we work with must be taken seriously. 20 It should be noted, however, that the WCC-ED study document is also very much aware of the need that the changes of religious landscapes in Europe and beyond also imply mutual challengers between faith language and secular mentalities: An interdisciplinary and dialectic approach implies critical communication between the two. The secular language will question the validity and range of the religious terms, and vice versa. 21 --- Recognizing the Value and Legitimacy of Different Social Shapes of Diaconia in World Christianity Another remarkable feature of the WCC-ED study document is its strong sense of a legitimate plurality of different forms of social ministries and diaconal action in World Christianity: There is no sense by which one special institutionalization of the diaconic, which for historical reasons has become dominant in one part of the world is regarded as normative or superior to other forms of social witness and diaconia. In Section 4.5. "Diaconia as discipleship" (WCC 2022, p. 48ff), there is a remarkable paragraph in which the different forms of diaconia are spelled out: In the history of the church, diaconia has always linked love for one's neighbors with the passion for justice and sustainability. It has become common understanding in world Christianity that this takes up different forms, which are

There has always been a conviction in the ecumenical movement that social service is part of the essence of what it means to be the church. The strengthening of social and political witness of churches through church-related agencies and networks in the context of the beginning development discourse from the 1960s until the 1990s, however, led to a scenario in which the secular language of development and political action received much attention. In 2014 at the Malawi consultation, the World Council of Churches (WCC) and Action of Churches Together (ACT Alliance) reviewed their working relationships and explored better collaboration between agencies and churches, stating that the specific nature of churches as entities with an added value and particular profile in the area of social and political service would be strengthened. From this consultation, a mandate was given by the WCC Central Committee to develop a study document on the understanding and core components of the concept of "ecumenical diaconia", which then was made available to the 11th Assembly of the WCC in Karlsruhe, 2022. This article introduces the context, major content, and key convictions of this study document and related discussions in Karlsruhe. It answers the question: what is the significance of the global ecumenical fellowship of churches in focusing on a common language of diaconia and a more explicit church-related terminology to describe and strengthen the social services of churches and their engagement in diaconia and social witness? It is recognized that this article is predominantly written from a German context and perspective in its construction of the international discourse on ecumenical diaconia.

Europe and beyond also imply mutual challengers between faith language and secular mentalities: An interdisciplinary and dialectic approach implies critical communication between the two. The secular language will question the validity and range of the religious terms, and vice versa. 21 --- Recognizing the Value and Legitimacy of Different Social Shapes of Diaconia in World Christianity Another remarkable feature of the WCC-ED study document is its strong sense of a legitimate plurality of different forms of social ministries and diaconal action in World Christianity: There is no sense by which one special institutionalization of the diaconic, which for historical reasons has become dominant in one part of the world is regarded as normative or superior to other forms of social witness and diaconia. In Section 4.5. "Diaconia as discipleship" (WCC 2022, p. 48ff), there is a remarkable paragraph in which the different forms of diaconia are spelled out: In the history of the church, diaconia has always linked love for one's neighbors with the passion for justice and sustainability. It has become common understanding in world Christianity that this takes up different forms, which are complementary and cannot be played off against each other. These different dimensions of the mandate for a diaconal church and for diaconal discipleship are not expressed in equal manner in all circumstances, but their realization depends on the cultural, political and institutional contexts of different churches. Thus, there exist different forms of diaconia, each with a different degree of professionalization and institutionalization: The intention of the WCC/ACT Alliance study document is clear here not to take one model of diaconia, for instance, the agency model that emerged in Scandinavian or German Christianity, as a priority model for other regions that have completely different political and social frameworks for their forms of institutionalization of social services. All different social models of diaconia are in themselves legitimate expressions of the one solid concern of the global churches to serve those in need and to stand up for the rights of the suffering and neglected or violated. --- Remaining Self-Determined and Distinct Church Priorities for Ecumenical Diaconia Being drawn into the international discourse on ecumenical diaconia as a churchbased development organization, the WCC-ED document sensitizes church-related actors to remain self-determined and independent in our decisions on priorities for projects and partnerships concerning ecumenical diaconal cooperation with churches in other parts of the world. Certainly, as an agency working within clearly regulated and mutually trustful relationships with state partners in the German context, church actors have to be attentive to regulations and frameworks regarding those who deserve primary attention and support. In this sense, church actors follow with attention and sensitivity the decisions concerning the groupings and prioritization of countries in the list of DAC countries, which, according to the Development Assistance Committee (DAC), are entitled to receive official development assistance (ODA) means. 22 It thus has an effect on the decisions of Christian agencies if countries are grouped in the list of least-developed countries, low-income countries, lowermiddle-income countries and higher-middle-income countries. 23 However, the decisions of the state regarding which countries are entitled to receive development funds cannot be the only criterion for church-based development actors to decide on how to prioritize their partners in the Global South. Current decisions on whether and to what extent to relate to partners, such as those in China, Brazil, South Africa, Ukraine, Russia, North Korea, or in other countries with high restrictions on civil society, are more complex than just to refer to economic indicators from the Organisation for Economic Co-operation and Development (OECD). An ecclesial assessment of the needs to support churches and churchrelated civil society organizations even in contexts that politically or economically would be regarded as not a priority country for state development cooperation should complement any political assessment and have its own ecumenical right and perspective. Maintaining ecumenical partnerships in ecumenical diaconia remains of strategic importance, even in contexts where, for one reason or another, no proper project funding currently is possible or advisable such as in closed countries with military governments or autocratic regimes. How to keep up with ecumenical diaconia in circumstances that deal with closed or highly restricted societies is a very sensitive issue in future collaboration in ecumenical diaconia as credibility over against other churches, credibility with regard to back donors, and effectives and protection against attempts to be politically misused are at stake. To keep up with small attempts, at least for transformative partnership relations and with efforts for capacity building and conceptual dialogue in issues of ecumenical social and peace ethics, remains of strategic significance, even in countries that remain difficult, partly closed, problematic, or completely blocked terrain for western or ecumenical Christian agencies. --- Strengthening the Collaboration between NCCs and ACT National Forums It was the main purpose of the consultation between the WCC and ACT Alliance in Malawi in 2014 that lead to the formulation of the ecumenical diaconia study document to prevent ecumenical platforms of churches from falling away from platforms of common action in humanitarian aid and development cooperation as established and continued within the structure of ACT Alliance. Explored through the work of the community of Practise on Religion and Development (CoPRD) and encouraged through the last ACT Assembly in Uppsala 2018 24 as well as by the WCC, leadership pilot projects have been undertaken in Malawi and Cameroon in order to find out how collaboration between National ACT Forums and National Council of Churches (NCCs) can be strengthened and improved in order to increase synergies between development actors and church actors on the ground. 25 That National ACT forums are mainly composed of representatives of northern development agencies cannot continue. 26 This also has an inbuilt and potential reformation dynamic relevance for the churches themselves and their democratic standards: the WCC/ACT Alliance document states: Agencies should be encouraged to develop strategies for working more closely with churches. Professor Christoph Stueckelberger, the founder of Globethics.net, claims that churches can often play a broader reaching role than single NGOs in transformation processes. He therefore recommends that agencies support processes that envisage strengthening democratic practices in the churches, developing responsible leadership, building competence, and creating corruption-free churches. (WCC 2022, p. 83) The interest in the localization and decolonialization of humanitarian aid as well as for the bridging of gaps between international funding mechanisms and local faith actors on the ground has been strong, particularly for learning achievements in relation to the challenges of the COVID-19 pandemic. In many cases during the pandemic, the local faith actors were to the most vivid and effective actors in providing education, medical training, and equipment for medical protection and assistance to those suffering from the pandemic. One aspect, which was mentioned several times within the working group on ecumenical diaconia in this regard, was the interest to move toward a more permanent think tank and collaborative coordination structure between the ACT Alliance and the WCC in the areas of theology, ethics, and pilot projects (practice) of ecumenical diaconia. The idea of a joint reference group on theology, training, and pilot projects for ecumenical diaconia to be called together by both the ACT Alliance and the WCC unit on public witness and ecumenical diaconia could potentially answer the need for both a strong commitment to ecumenical social ethics and nurturing networks of social action and diaconia on foundational and ethical concepts and related pilot projects that inform the current work of the global fellowship of churches in the area of ecumenical diaconia. The formation of a joint reference group on ecumenical dianoia was finally included in the decisions and recommendations made by the WCC Assembly in Karlsruhe through the report of its Programme Guidelines Committee. 27 --- Building Bridges with Churches That Do Not (Yet) Have a Long Tradition of Ecumenical Social Thinking and Diaconia The WCC-ED study document is quite clear on the conviction that the unity of churches' commitment in ecumenical diaconia is not a finished product but an openended process and always needs to be strengthened and reconstructed. There still is a lack of simultaneity between different confessional traditions and contextual approaches to Christian social witness and prophetic diaconia. There are churches both in the membership of the WCC as well as outside that, for several reasons, have had a shorter history only with regard to involvement in public social witness, ecumenical diaconia, and advocacy for issues of justice, peace, and integrity of creation because their social, political, and doctrinal background did not allow for making regular advances in this regard. One of the main intentions of the WCC-ED document is the closing of gaps in this regard and the initiation of creative mutual leaning processes between churches of different cultural and social contexts. There are many churches in World Christianity on the local level that have a genuine and vital concern for social witness and ecological care but, at the same time, did not develop capacities to build up professional regional institutions of diaconal action for establishing partnerships with churches in different geographical contexts or institutes for ecumenical social ethics and dialogue between faith and social sciences on issues that need to inform the public witness of churches today. 28 The processes of internationalization in the area of diaconia are moving quickly and rapidly changing the landscapes of World Christianity. However, very few people know or are aware of the fact that some churches and their social service agencies in Nigeria or China are much bigger than in churches in Germany or Switzerland today. We need a more deliberate process of ecumenization and internationalization of diaconia today after 175 years of national diaconia in Germany. How to live and deal with different speeds, dialects, and contexts of social diaconia actors within the global Christian family is one of the key challenges particularly in the current situation. Rather than co-opting a political language of swift exclusion, rejection, and denunciation for those churches who are different than our own traditional patterns of performing diaconia, we need to become open for new players, for instance, with churches with a migration background. 29 We need to be careful to not just applaud our given standards and ourselves but instead install mechanisms of mutual intercontextual learning from new language and innovative practices of other churches in World Christianity. Keeping the gates open for building new leadership in public theology, for responsible peace ethics, and for a common work on social and ecological diaconia in the end might provide better results rather than a trend of excluding those churches that currently are still lacking a clear, responsible, and focused statement on social ethics and diaconia. It is thus deeply tragic that the current war in Ukraine has not only gravely damaged the relationships between Russia and Ukraine, the east and the west, as well as between Orthodoxy and Protestantism in general: it also has diminished the open-minded learning interests of several Russian Orthodox theologians to learn from achievements and models of diaconial work in western churches (and vice versa), which was a major opportunity for open dialogue between Orthodoxy and Protestantism that now seems disabled and might remain closed for some time (Werner 2020). Much remains to be achieved here. --- Deepening the Common Commitment to Training and Theological Education for Ecumenical Diaconia The WCC-ED study document is a strong reminder that competence building for diaconal action and presence in fields of social needs as well as in environmental stewardship is an essential priority in all churches, as the world is confronted with huge challenges of social inequality, long-term impacts of climate change, and destruction of natural biodiversity. Achieving the core goals of the SDG agenda will not be possible without involving the voice and services of churches and religious communities that have a mobilizing capacity beyond national governments. Demand for highly competent experts in diaconal services, poverty education, and ecological transformation will certainly be on the rise in the years to come. State governments should provide platforms of mutual collaboration and financial support with churches and other religious actors to realize their potential for eco-social transformation. It would be worthwhile for the WCC or regional actors of the ACT Alliance to convene a summit on training for diaconia and social services. The WCC and ACT could further deepen their joint vision for training and competence building for diaconia, social witness, and eco-diaconia while supporting the promotion of the closer collaboration of churches on these issues. Several sections in the last chapter of the WCC/ACT document are related to training and strengthening networks of ecumenical action in diaconia. 30 The WCC document indicates a vast variety and different needs tor training for diaconia: Training for diaconia in local churches, specialized ministries and agencies is currently developed at very different levels depending on the history, socio-political circumstances and structural frameworks available for civil society organizations. In some churches, organization for diaconal institutions was not possible at all (for instance during the period of the Soviet Union). In other churches there are small scale diaconal activities possible within local churches, but no larger diaconal institutions yet. Some churches are in circumstances where it is possible to receive partial state support for Christian social diaconal services rendered on behalf of the churches, in other settings this is not possible at all. These differences are mirrored by very different levels of diaconal training and education for Christian social services. There are settings with strong and established schools, colleges or universities of applied sciences for diaconal sciences and social work, whereas in other settings there are almost no institutes of training and education available. (WCC 2022, p. 113) The optimal follow-up project for the current WCC-ED study document would be a small research project to be pursued by a significant research institute to produce an advanced and diversified list with recognized and qualified training institutions of churches for ecumenical diaconia, prophetic diaconia, diaconia management, and ecological diakonia, which could be recommended to churches and agencies for increased mutual collaboration in training and research. This should be accompanied by considerations to revive a proper scholarship program, which is essential for building up competence in diaconal leadership for churches in the Global South and the East. Churches studying the WCC-ED document together should closely reflect on a strategic suggestion, which is given in this regard in the same section: Recent years have however seen a significant rise of interests and motivations for training capacities and curriculum model development in diaconia studies from churches in various contexts. It belongs to the core tasks of the ecumenical movement in the 21st century to enrich and facilitate diaconal learning and education in as many churches and different settings possible. This will allow each local church to find access to appropriate training resources for strengthening diaconal capacity and to engage in mutual exchange, cross-border education, and ecumenical partnership programs for diaconal learning, which demands a common major commitment of diaconal agencies, development agencies, mission agencies, and educational networks. This challenges churches, agencies, and ecumenical bodies to: --- • Include capacity building in diaconia in their strategic and programmatic plans; Producing diaconal and developmental capacity building will also include a substantial reconsideration of the priorities of development agencies, with the goal of increased funding to be made available for study, training, and scholarships projects in this area. There are promising institutes that provide an international training and research program on ecumenical diaconia such as the Institute for Diaconal Management Studies (IDM Institute) in Bielefeld, 31 which is partly funded by Bread for the World and UEM. I hope that it can continue its work. --- Revitalizing the Common Memory of Ecumenical Social Ethics It might belong to the role of those who have grown older to remind us of the value of those who have worked and witnessed before us in the cloud of witnesses and the treasures of wisdom, vision, and ecumenical tradition that they have given to us. These treasures of ecumenical wisdom and knowledge, particularly in the area of ecumenical social ethics, should not just stay with us personally but should stay with our organizations and networks that we have worked with despite the natural flow of different generations and staff. How to maintain an intergenerational management of ecumenical and diaconal knowledge in social organizations and churches is a challenge to the whole of the ecumenical movement but also specifically to Christian development organizations and church-related social services, and this applies to churches both in the Global North as well as in the Global South. How do we avoid the loss of ecumenical memory weakening the future witness and service of churches in the global ecumenical fellowship? How do we secure a smooth transition to future generations without keeping specialized knowledge as a personal property clinging only to well-established and closed communication circles? Each Christian church and organization will have to answer these questions in one way or another. With shrinking state-and church-related funds made available, Christian diaconia and development cooperation will also have to be more creative to revive the ancient culture of charitable donations and philanthropy in order to answer the rising needs for humanitarian and social development assistance. How to develop proper standards for accountability and proper ethical standards for private actors in this field is a debate worth continuing without ideological one-sidedness or naiveté. We should be encouraged to learn again that churches in previous centuries and in different regional contexts have developed impressive institutions of diaconal assistance, not just in in charity approaches but also for political advocacy and justice concerns. To be explorative in terms of new models of financial sharing and new mechanisms of ecumenical sharing of resources in ecumenical diaconia is a priority for the next decades, as indicated in chapter 6.6. on diaconal assets in the WCC ED paper (WCC 2022, p. 73ff). --- Keeping Together Social and Ecological Diaconia Economic inequality, climate change, and its effects on the health situation of both humanity and the realm of animals as well as the global recession of democracy belong to the core challenges of the 21st century, which any involvement of churches in social diaconia must address. Keeping together social diakonia, which articulates practical answers to the growing economic inequality and exclusivist tendencies of autocratic governments as well as eco-diakonia, which deals with the destruction of our planet and its natural resources, which cannot be seen anymore as exclusive or antagonistic mandates: they belong together inseparably. Experts have shown that the depletion of natural resources is accelerating each year. This is made visible by the preponing of the date of the so-called "earth overshoot day": the projected date in the year when our current economic system has exhausted all its reservoirs in terms of natural resources for which it is entitled if this is balanced with the regenerative capacities of this Earth. For Germany, this earth overshoot day was projected in 2022 as having occurred already on 4 May, for humanity as a whole on 3 August, for South Africa on 1 June. 32 All of the resources, which proportionately could be used and absorbed this year within the given planetary boundaries, are gone by this day: any continuation of the same rates of resources' depletion is simply living at the expense of future generations, i.e., living on credit. As the introduction of binding regulations for the upper limits related to resources extraction would need to be obligatory for all; this implies a shrinking economy instead of a continuously growing economy, and this will accelerate the social conflicts between different income groups. How to reduce global resources extraction while not fueling global income distribution and social subsidies conflicts will be a core challenge for steering a course of a just ecological transition while being the key theme of ecumenical diaconia. Keeping the issues of justice and overcoming of social inequality and exclusion as well as issues of the ecological transition to a new circular and non-fossil economy on the agenda with paramount priority will be the challenge to remain for decades on the agenda of both ACT Alliance and the WCC. How to hold together social justice and protection of ecological resources? How to develop both models of sustainable diaconia and of inclusive communities will remain the key task for the years ahead, which is the period of the biggest transition of humankind since industrialization. It was only in October 2021 that diaconia Germany agreed and decided on guidelines to make "sustainability" (Nachhaltigkeit) a key criterion for all its operational dimensions and implementation of this is still underway in its different layers. 33 Both issues, the transition to sustainable models of development as well as the overcoming of growing social inequality, are the key thematic areas that have been explored and laid out in the contributions of two international volumes for training in ecumenical diakonia, which were produced for the WCC Assembly in Karlsruhe in September 2022 and have been presented to its Ecumenical Conversation on Ecumenical Diaconia, namely the twin publications of the International Handbook on Ecumenical Diaconia (Ampony et al. 2021) and the International Handbook on Creation Care and Eco Diaconia (Beros et al. 2022). 34 --- Conclusion: Strengthening Ecumenical Diaconia as a Common Christian Response to Three Major Challenges of the 21st Century To summarize, we have three key observations from these reflections: (a) The joint study document of the WCC and ACT alliance represents a unique chance to intensify the ongoing global learning and dialogue process on the relationship between churches and diaconia, which can involve all segments of World Christianity, including mainline churches, orthodox, evangelical, and charismatic churches, which potentially can become a push for renewing social and ecological awareness and witness of churches globally, in other words to reformulate something like what Social Protestantism has been for some of the western churches in Europe and the North Atlantic. It presents a new step in the development of ecumenical diaconia because it abandons the individualistic concept of diaconia as individual acts of mercy and seeks ways to help people in need, which is characteristic of all churches, though diverse in shape and context. The transformative potential of social and ecological diaconia should be much more intentionally realized by churches, development agencies, and national governments. (b) The ecumenical discourse on ecumenical diaconia presents a challenge to deepen and rediscover the integrity and comprehensiveness of the diaconal approach of Christian churches and to more visibly relate to each other aspects of social diaconia and ecological diaconia in a context where the SDG agenda is far from being likely to be fulfilled and completed by 2030 (WCC. EC 10 Report 2022). 35 It has been a major learning process in the history of World Christianity to move from individual-charitybased diaconia to structural diaconia, political lobbying work, and international social ethics, but it is still a major new learning process to move from social and political diaconia to ecological diaconia today. How to spell out ecological sustainability ethics for diaconal agencies and services to integrate the diversity of diaconal actors (i.e., churches with migration or ethnic and confessional minority backgrounds) into the national networks of engagement for social and ecological diaconia remains a key mandate for the future. The plea to focus on a massive new engagement for ecumenical diaconia in World Christianity is also an answer to the unprecedented intensification and accumulation of the global poverty, climate, hunger, health, and militarization crises, which have led to the fact that, for the first time in the history of the Human Development Index (HDI), trends have been moving backward on a global scale at an alarming rate for the last 2-3 years. 36 Thus, the strengthening ecumenical diaconia also presents a common Christian response to the three major challenges of the 21st century: rising economic inequality, the climate and health crises, and global recession of democracy. (c) In a context where church membership seems to be shrinking and social as well as humanitarian needs and ecological disasters seem to be dramatically increasing, the integration of churches and social and ecological diaconia is also a vital contribution to strengthen the visibility, outreach, and relevance of Christian churches both in western as well as in nonwestern societies. It is far from being determined what is meant by diaconia remaining the "other and second language" of the Gospel, which is equal in theological significance and communicative reach to the language of proclamation and sermons. See: https://www.cevaa.org/ (accessed on 10 January 2023). --- 8 ACT Alliance is composed of more than 140 faith-based member organisations working in long-term development, advocacy and humanitarian assistance. See: https://actalliance.org/ (accessed on 10 January 2023). 9 The WCC is a fellowship of 352 member churches who together represent more than half a billion Christians around the world. 10 See LWF consultation on the future of theological education in Wittenberg 2012: https://www.lutheranworld.org/news/lwfgeneral-secretary-calls-greater-role-diaconia-and-advocacy-theological-formation (accessed on 10 January 2023). 11 See particularly the work on Diaconia in the IECLB in Brazil from which the later General Secretary of ACT Alliance Rudelmar Bueno de Fario emerged. See: https://www.oikoumene.org/de/news/new-act-general-secretary-envisions-more-propheticdiaconia (accessed on 10 January 2023). 12 See: https://www.oikoumene.org/en/press-centre/news/church-organizations-explore-ways-of-strengthening-relationships (accessed on 10 January 2023). See: https://www.oikoumene.org/news/archbishop-antje-jackelen-opens-act-assembly-in-uppsala (accessed on 10 January 2023). 25 See: https://actalliance.org/act-assembly-posts/story-of-hope-wcc-act-alliance-pilot-learning-process-in-malawi-and-cameroonto-strengthen-collaboration/ (accessed on 10 January 2023); also in: https://www.oikoumene.org/news/wcc-act-alliance-pilotlearning-process-in-malawi-and-cameroon-to-strengthen-collaboration (accessed on 10 January 2023). --- 26 See Fidon Mwombeki from AACC in a statement from 2022: "We face a problem of structural asymmetry between European and Southern churches, in the sense that in the South there are no strong "specialized ministries" like in the north. Even where there are, they have been formed as a result of encouragement, or demand from the north, for the churches to create similar structures like Europe. So, several "development services" have been established by some churches in the South which till now cause some frictions with the local churches. This asymmetry has created challenges for ACT Alliance, whose "majority" who are technically churches in the South, many do not even know or remember they are members of ACT Alliance. For that reason, the national ACT Forums in several countries are basically formed by national offices of Northern based specialized ministries, with minimal involvement of the national churches who have no separate "ministries." We must act quickly address this asymmetry to achieve the justifiable aspiration of ACT Alliance. --- Data Availability Statement: Not applicable. --- Conflicts of Interest: The author declares no conflict of interest. --- Notes 1 See: https://www.oikoumene.org/news/major-wcc-and-act-publication-on-ecumenical-diaconia-now-available-in-german; all language versions available on: https://www.oikoumene.org/de/resources/publications/ecumenical-diaconia (accessed on 10 January 2023). See: https://www.vemission.org/en/ (accessed on 10 January 2023).

There has always been a conviction in the ecumenical movement that social service is part of the essence of what it means to be the church. The strengthening of social and political witness of churches through church-related agencies and networks in the context of the beginning development discourse from the 1960s until the 1990s, however, led to a scenario in which the secular language of development and political action received much attention. In 2014 at the Malawi consultation, the World Council of Churches (WCC) and Action of Churches Together (ACT Alliance) reviewed their working relationships and explored better collaboration between agencies and churches, stating that the specific nature of churches as entities with an added value and particular profile in the area of social and political service would be strengthened. From this consultation, a mandate was given by the WCC Central Committee to develop a study document on the understanding and core components of the concept of "ecumenical diaconia", which then was made available to the 11th Assembly of the WCC in Karlsruhe, 2022. This article introduces the context, major content, and key convictions of this study document and related discussions in Karlsruhe. It answers the question: what is the significance of the global ecumenical fellowship of churches in focusing on a common language of diaconia and a more explicit church-related terminology to describe and strengthen the social services of churches and their engagement in diaconia and social witness? It is recognized that this article is predominantly written from a German context and perspective in its construction of the international discourse on ecumenical diaconia.

Associations between depressive symptoms and 5-year subsequent work nonparticipation due to long-term sickness absence, unemployment and early retirement in a cohort of 2,413 employees in Germany Uwe Rose 1*, Norbert Kersten 1, Dagmar Pattloch 1, Paul Maurice Conway 2 and Hermann Burr 1 --- Background This paper deals with associations between depressive symptoms and work nonparticipation. The opposite of work nonparticipation has been conceptualized as " [...] the capability and/or opportunity to participate in the workforce, fulfilling one's work role" [1]. Over the last two decades, the proportion of registered sickness absence events attributed to mental health problems has increased in Germany [2]. Absence spells lasting less than six weeks account for approximately half of all days of absence due to sickness [3]. The proportion of disability pensions due to mental disorders has also increased from 19 to 43% over 22 years [4]. These figures point to the substantial role that mental health plays in terms of individual (reduced income) as well as direct (health problems) and indirect (taxes, support payments) costs for the social security system [5]. Mental health impairments can be linked to both short-and long-term forms of work nonparticipation. While sickness absence is typically a temporary state, early retirement (ER) and disability pension are -with a few exceptions -permanent [6]. In addition to sickness absence and ER before statutory retirement, unemployment (UE) should be regarded as a further type of work nonparticipation [6,7]. Both temporary and permanent states of work nonparticipation were previously found in association with productivity losses [5]. In most countries, cases of registered sickness absence and disability pension are attributed to one or more diagnoses. This attribution is the result of a decision-making process that takes not only clinical but also administrative and regulatory aspects into account [8,9]. Thus, the diagnoses reported in administrative registries as the underlying reason for sickness absence or disability pension might not necessarily align with the real associations of these diagnoses with sickness absence and disability pension. This is reflected, for example, by the increasing number of events of sickness absence and disability pension due to mental health problems that occurred in recent decades, which stands in contrast with the rather stable prevalence of mental disorders in Europe [10,11]. Therefore, it might be that administrative statistics shed light on the predictive role of mental health in relation to sickness absence and disability pension only to a limited degree. To bridge this gap, epidemiological evidence on the possible causal relationship between mental health conditions and work (non)participation is needed that relies on longitudinal studies examining depressive disorders or depressive symptoms as predictors of sickness absence, disability pension, ER, and UE. Among mental health conditions, depression is one of the most disabling [11]. Usually, a diagnostic interview is the basis for classifying an individual in a discrete depression category [12,13], whereas depression as a continuum is traditionally assessed by using symptom inventories administered via paper and pencil questionnaires [14,15]. To date, the evidence on depressive symptoms as a predictor of work (non)participation based on representative samples and cohort study designs is limited. Most existing studies examined events of work nonparticipation but did not take their duration into account. A few studies suggested an association between depressive symptoms and sickness absence. Specifically, the Household Income and Labour Dynamics (HILDA) Survey in Australia [16] and the Maastricht Cohort Study in The Netherlands [17] showed that increased levels of depressive symptoms were associated with higher incidence rate ratios or hazard rates of short-and long-term sickness absence. The rates were slightly higher among males in the first study, while there were no gender differences in the latter study. A similar result was observed in the Danish National Working Environment Survey (DANES), which linked self-reported depressive symptoms at baseline to 60-week follow-up data on long-term sickness (<unk> 4 weeks) derived from a nationwide register, with no stratification by gender [18]. However, long-term sickness absence is only one type of exit from paid work. In previous studies, depressive symptoms were also associated with forms of work nonparticipation other than sickness absence [19]. For example, in a study based on the Survey of Health, Ageing and Retirement in Europe (SHARE), depressive symptoms were associated with disability benefits and UE [20]. For males and females with depressive symptoms, the hazard rates for disability benefits were HR = 3.29 (95% CI = 1.99-5.46) and HR = 2.10 (95% CI = 1.17-3.77), respectively, while the association between depressive symptoms and UE was not significant (males: HR = 1.55 (95%-CI = 0.94-2.57); females: HR = 1.14; 95%-CI = 0.73-1.76). In their systematic review on the link between poor health and exit from paid work, Rijn et al. [19] found associations between indicators of poor (mental) health and later disability pension, UE and ER. In sum, the current state of the art suggests that, with only a few exceptions [20,21], there is to date limited knowledge about the association between depressive symptoms and different types of work nonparticipation. This limits the current understanding of whether the association between depressive symptoms and long-term sickness absence (LTSA; i.e., the most frequently investigated outcome) [16-18, 21, 22] is of a similar or a different size than their association with UE [20,21] or ER [17,23]. For instance, in health economic analyses this knowledge gap prevents one from summarising different types of work nonparticipation into single measures. Events of work nonparticipation can be recorded both in terms of occurrence and their duration. However, to date longitudinal studies have predominantly captured event occurrence by treating events as dichotomous outcomes. While event-related analysis is important for examining the general causal relationship between depressive symptoms and different outcomes, such as permanent exit from paid employment, it does not capture its duration, which allows for the assessment of the associated productivity loss. Only a few studies have previously attempted to model these time aspects [17,21,[24][25][26]. In addition, while most available studies have stratified the analyses by gender [16,17,[20][21][22], only one [17] found significant indications of gender differences in the association between mental health and work nonparticipation. However, none of these studies employed statistical tests to examine interaction by gender. Based on the above-mentioned considerations, in the present study we aim to investigate the associations of depressive symptoms with two aspects of work nonparticipation, that is, occurrence and duration of events of work nonparticipation. Based on the aforementioned literature review [20,21], we do assume that the associations between depressive symptoms and work participation will differ between males and females. --- Methods --- Sample We analysed a cohort based on the Study on Mental Health at Work (in German: S-MGA), using both the baseline (2011/12) and follow-up (2017) measurements. This study was initiated by the Federal Institute of Occupational Safety and Health in collaboration with the Institute of Employment Research (IAB). The latter incorporates data on all employees in Germany subject to social security contributions in the Integrated Employment Biographies (IEB) register, which covers more than 80% of the working population in Germany. Employees subject to social insurance contributions as of 31 December 2010 were registered in the IEB; civil servants, selfemployed individuals and freelancers were excluded by this definition. The S-MGA includes cohorts of individuals born between 1951 and 1980. To facilitate face-to-face interviews, a two-stage cluster sampling procedure was applied to a random sample of 206 municipalities in Germany proportionately stratified by region and population, followed by a random gross sample of 13,590 addresses drawn within selected municipalities [27]. A flow-chart of participation is shown in Fig. 1. Altogether, 4511 individuals participated in a computer-assisted personal interview (CAPI) at baseline. The follow-up took place in 2017 (n = 2640). Since there was a gap of an average of 13 months between the sampling and the interview dates, in which participants could retire, 66 participants were excluded from the analysis. In addition, 500 participants from the baseline measurement were excluded because they presented missing values on depressive symptoms or skill level (see the subsections 'Independent variables' and 'Covariates' below). As for the cohort, the largest loss was due to 1509 respondents not taking part in the follow-up survey after having participated at baseline, and to 23 respondents for whom the employment biography could not be established at follow-up. The final sample included in the present analyses consisted of a cohort of 2413 individuals (n = 1168 males and n = 1245 females) that participate in both measurements. This corresponds to a participation in the cohort study of 18% (Fig. 1). Based on comparisons with the sampling frame of the study, participation in the cohort was independent of gender, while it was somewhat associated with age (20% among those aged 55-60, 15% among those aged [31][32][33][34][35][36] and skill level (24% among professionals and managers and 14% among unskilled workers) [28]. All information was obtained in the respondents' home, mainly through the CAPI, whereas sensitive information about depressive symptoms was gathered through paperand-pencil questionnaires that were completed without the interviewer being present. --- Independent variables We categorized depressive symptoms into three categories by means of the Patient Health Questionnaire (PHQ-9) [29]:'medium to severe' (<unk> 10) [30],'mild' (5-9), and 'none' (0-4). The PHQ-9 yields a sum score based on nine items which were preceded by the following question: 'Over the last 2 weeks, how often have you been bothered by any of the following problems?<unk> The nine items were: 'Little interest or pleasure in doing things', 'Feeling down, depressed or hopeless' 'Difficulty falling asleep or sleeping or increased sleep', 'Tiredness or feeling unable to have energy', 'Decreased appetite or excessive need to eat', 'Bad opinion of yourself ', 'Difficulty concentrating on something', 'Slowed speech/movement or restlessness ("fidgety")', 'Thoughts that you would rather be dead or want to self-inflict pain'. The response options were: 'Not at all' (0), 'Several days' (1), 'More than half the days' (2) and 'Nearly every day' (3). The score for depressive symptoms is computed as the sum of all 9 items. Cronbach's Alpha was.82 and the inter-item correlations were between.22 and.51. Based on previous analyses performed by authors of this study, the distribution of the PHQ values was positively skewed, which is due to our baseline sample being healthier than the general population in the same age range. --- Outcomes We considered different forms of work nonparticipation, that is, LTSA, UE and ER, occurring between baseline and follow-up. A few events of work nonparticipation for which participants did not report the specific type were also considered but not examined separately due to their low incidence (Table 1). Each event was measured in terms of the year and month it started and ended. Regarding LTSA, respondents were asked to only report events lasting at least 6 weeks. This restriction did not apply to UE and ER. As in our analyses we treated the various forms of work nonparticipation as mutually exclusive events, we considered only one event as reported by participants in a given year and month. However, events could take place consecutively over time; for instance, first an event of sickness absence followed by an event of ER. Since respondents were allowed to report LTSA events which overlapped with events of UE and ER, for the present analyses we prioritized UE and ER, respectively, over LTSA. Therefore, any events of LTSA that overlapped with UE and ER were not considered. We also calculated an index, labelled total work nonparticipation (TWNP), that included all events of LTSA, ER, and UE, and other types of work nonparticipation that could not be ascribed to any of these forms. --- Covariates We considered gender, age at baseline, skill level and employment status as covariates. In line with previous studies [16,17,[20][21][22], we used gender as stratification variable. Age at baseline was categorized into three decades (31-40, 41-50, 51-60). The skill level of the job was based on the International Standard Classification of Occupations (ISCO 08) of the International Labour Office and categorized into four groups: unskilled workers (ISCO groups 9), skilled workers (4-8), semi-professionals (3), and professionals (1,2). In the analyses, the most frequent group ('skilled workers') was used as reference. The variable "in paid employment at baseline interview" was treated as dichotomous ('no' vs. 'yes'). We did not include indicators such as income and education as covariates. We did so because in our data skill level correlated with both education (Kendall Tau b = 0.48) and income (categorized in quintiles; Kendall Tau b =0. 35) and introducing them in the analyses would have led to less stable regression coefficients. We prioritized skill level as education was a less proximal indicator of --- Statistical analysis The number of months as outcome is a semicontinuous variable that takes exclusively non-negative integer values and is zero-inflated (i.e., a large stack of data points concentrates around zero as the lower bound). Therefore, the distribution appears as a mixture of zeros and positively skewed non-zero values. A common regression analysis or multiple regression analysis cannot be applied to such a distribution because assumptions such as normality and homoscedasticity are not met. Thus, we decided to use two-part models to analyse the present data. The likelihood function l of a two-part model [31, 32], with a variance parameter <unk>, takes the following form: l (<unk> 1, <unk> 2, <unk>) = <unk> <unk> <unk>i:yi=0<unk> P (yi = 0|x 1,i, <unk> 1 ) <unk> <unk> <unk> <unk> <unk>i:yi>0<unk> P (yi > 0|x 1,i, <unk> 1 )f ( yi|x 2,i, <unk> 2, <unk>) <unk> <unk> <unk> 1 and <unk> 2 are two sets of regression parameters to be estimated in the model. f is a density function specified in the concrete case. This likelihood function has only one global maximum, which can be decomposed into two parts, each of which can be maximized separately. First, the logistic regression part with parameters <unk> 1 was used to model zero and non-zero values. Second, conditional on all non-zero values, generalized linear models (GLMs) with a gamma distribution and a log-link function were performed to model the non-zero values with parameter <unk> 2. In other words, the logistic regression component of the two-part model was used to model the event-related data. Odds ratios (OR) were calculated to estimate the risk of at least one event of nonparticipation between baseline and follow-up. Conditional on all non-zero events, GLM is used to model the number of months of work nonparticipation expected for these events. The estimates of a GLM with a logarithmic link function are expressed in terms of Exp(<unk>), which is similar to an Odds Ratio (OR) derived from a logistic regression. Exp(<unk>) and OR operate as multiplier. The predictive value, i.e., the number of months, is calculated by multiplying the exponential function of the intercept (Exp (<unk> 0 )) by the exponential functions of the regression coefficients. In the GLM analyses, we only considered participants who reported at least one event of work nonparticipation. As we expected that depressive symptoms were conditional of gender, in an additional analysis pooling together males and females, we tested if gender interacted with depressive symptoms by entering an interaction term (depressive symptoms*gender) in both the logistic and GLM regressions. In all analyses, we calculated both crude estimates and estimates adjusted for the covariates. All statistical analyses were performed using IBM® SPSS® Statistics, version 27.01.0. --- Results --- Sample description Tables 2 and1 show the sample characteristics in both genders. The total follow-up time was 73,079 months for males and 77,913 months for females. The mean followup time was 63 months in both genders. --- Depressive symptoms and events of work nonparticipation Tables 3 and4 show, for males and females respectively, the OR adjusted for age, skill level, UE at baseline and the corresponding 95% confidence intervals (95% CI). Among males, medium to severe depressive symptoms (OR = 3.22; 95% CI = 1.90-5.50) mild symptoms (OR = 1.59; 95% CI = 1.19-2.11) were significantly associated with TWNP. Both mild depressive symptoms (OR = 1.48; 95% CI 1.08-2.02) and medium to severe depressive symptoms (OR = 2.96; 95% CI = 1.76-4.97) were associated with LTSA. Medium to severe depressive symptoms were also associated with UE (OR = 5.17; 95% CI = 2.58-10.36). Mild depressive symptoms were significantly associated with ER (OR = 1.72; 95% CI = 1.01-2.92), while medium to severe symptoms were not. Among females, medium to severe depressive symptoms were also associated with TWNP (OR = 1.92; 95% CI = 1.29-2.87) and with LTSA (OR = 2.57; 95% CI = 1.69-3.91). There was no significant association between depressive symptoms and both UE and ER. --- Depressive symptoms and duration of work nonparticipation The second column in Tables 5 and6 displays, for males and females respectively, the estimated number of months among those with no depressive symptoms (PHQ score 0-4) (Exp (<unk> 0 )). Among males, there was a significant association between mild depressive symp- Among females, the duration of TWNP was more than a third longer [Exp(<unk>) = 1.38; 95% CI = 1.05-1.83] among participants presenting medium to severe depressive symptoms. The duration of LTSA was longer in both depressive symptoms groups [Mild: Exp(<unk>) = 1.30; 95% CI = 1.07-1.57 -Medium to severe: Exp(<unk>) = 1.49; 95% CI = 1.15-1.94]. Medium to severe depressive symptoms was not associated significantly with duration of ER and UE. --- Interactions by gender Tests of interaction showed that, as a whole, the associations between depressive symptoms and events of work nonparticipation did not differ in the two genders. The only exception is the association between medium to severe depressive symptoms and UE, which was stronger among males than among females (p for interaction = 0.008; Males: OR = 5.17, 95% CI = 2.58-10.36; Females: OR = 1.13, 95% CI = 0.57-2.25). Tests of interaction showed no gender differences in the associations between depressive symptoms and duration of work nonparticipation. --- Discussion The present study suggests that medium to severe depressive symptoms are associated with subsequent TWNP, both in terms of event occurrence and duration. In addition, we found indications that also mild depressive symptoms are associated with subsequent TWNP, at least when it comes to the occurrence of events. The present study suggests also significant associations between depressive symptoms and subsequent LTSA and UE. However, the associations with ER were not significant. With one exception only (i.e., events of UE, for which the associations were stronger among males), we did not find support for our expectation that the association of depressive symptoms with different forms of work nonparticipation would depend on gender. --- Comparison with other studies To our knowledge, only a few previous longitudinal studies based on representative samples of employees have hitherto investigated associations between self-reported depressive symptoms and subsequent work nonparticipation [16-18, 20-22, 33]. Most of these studies were focused on sickness absence or LTSA [16-18, 21, 22]; two studies examined UE [20,21], while another two focused on disability pension [20,33]. Only one study used aggregated measures of work nonparticipation [21]. Most existing studies were conducted in Denmark, except for the SHARE study, which included a number of countries (mainly European) and was based on a sample of older employees [20], and two studies from Australia (HILDA) and the Netherlands (Maastricht cohort study) [16,17]. Regarding TWNP, the aforementioned Danish study confirmed our finding that depressive symptoms were associated with a longer duration of TWNP, but only limited to LTSA and UE [21]. In relation to LTSA, our study confirmed previous findings from Australia [16], the Netherlands [17] and Denmark [18,21,22], all suggesting significant associations, similar in strength to the ones we observed, between depressive symptoms and LTSA [16-18, 21, 22]. Moreover, our findings are in line with two previous studies showing associations between mild depressive symptoms and LTSA [16,18]. Two further studies confirmed the findings of the present study by suggesting a link between depressive symptoms and a longer duration of sickness absence [16,21]. As for UE, our findings are in line with the aforementioned SHARE study and one of the mentioned studies from Denmark [20,21]. We could not find any previous studies examining ER while combining disability pension and other types of ER as we did here. Two studies found associations with disability pension [20,33], with hazard ratios indicating a risk that was two to three times higher among those with depressive symptoms. As a whole, the limited available evidence about the possible associations between depressive symptoms and subsequent work nonparticipation makes it difficult to establish if the German situation deviates from that of other countries. In addition, measurements of depressive symptoms differ in the available studies. The instruments employed include the Major Depression Inventory (MDI) [18,21], the Mental Health Inventory (MH-5) [16,18,22,33], the EURO-D and the Hospital Anxiety and Depression scale HAD-D [17]. The question as to whether associations with work nonparticipation change as a function of how depression is measured is, therefore, still open. --- Strengths One strength of the present study lies in the use of a longitudinal design, with depressive symptoms measured at baseline and employment biography followed up over the subsequent five years. The prevalence of depressive symptoms was captured in a representative sample of the German working population. The comparison of the sample characteristics with the referent population indicated no major selection bias at baseline. Additional strengths are the use of the PHQ-9 to measure depressive symptoms. This instrument is well established both in Germany and in international representative studies. With a cut-off value of > 9, it offers a case classification; in addition, the sum score provides a quantitative measure of depressive symptoms. From the perspective of prevention, it is important to recognize depressive symptoms before the full-blown disorder manifests. Supporting this, the present findings suggest that already mild depressive symptoms are associated with TWNP and LTSA. A study comparing self-reports and register-based measures of LTSA, UE and ER over a longer time span suggests that self-reports are a valid tool to assess these phenomena [34]. Such data enable not only the recording of an event but also the quantification of its duration. --- Weaknesses A first limitation is that no firm causal conclusions can be drawn based on the present findings. Features of our study design such as the measurement of the predictors prior to the measurement of the outcomes, as well as the inclusion of confounders through stratification and adjustment, are useful in supporting casual interpretations with regards to the examined associations [35]. Despite these features can reduce sources of bias, the present study remains an observational one, which limits conclusions about causality. Second, participation in the cohort was low (18%). We did not observe differences in attrition due to gender, while differences due to age and skill level were moderate (see Method section). We believe that these differences in attrition did not introduce major biases in our risk estimates, considering that age and skill level were adjusted for in all analyses. A third weakness might be that the study addresses only longer work nonparticipation events, as only events lasting months, instead of weeks or days, were considered. Additionally, only sickness absence episodes of at least 6 weeks were investigated. According to reports by the German Federal Ministry of Health, more than half of all recorded sick days relate to events with a duration below 42 days [3]. This could have led to a substantial under-recording of the total amount of sickness absence in our study. In the German context, the use of shortterm sickness absence is limited as the German insurance system comprises a large number of different health insurance companies that use non comparable registration systems. As a result, we were unable to examine if depressive symptoms presented the same associations with subsequent short-term absence as with LTSA. Fourth, the present study only considered the pathway from depressive symptoms to work nonparticipation, while reverse causation was not considered. In some cases, events of work nonparticipation such as UE might deteriorate mental health [36]; in other cases, events such as ER might relate to improved health [37]. Two-wave study designs are not sufficient to shed light on the processes, which require multi-wave studies [38]. Currently, the third wave of the S-MGA study is underway, which might provide a better basis for examining reversed causation. Fifth, we adjusted for skill level and not for education in the regression analyses (see the covariates subsection of the Method section). As skill level and education were strongly correlated, adjusting for both could have led to unstable regression coefficients in light of the limited sample size as well as the low prevalence of depressive symptoms (Tables 2 and1). --- Conclusions From a clinical point of view, our study suggests that medium to severe depressive symptoms can lead to increased work nonparticipation. In addition, even mildlevel depression can be predictive of increased work nonparticipation. Therefore, the early detection of possible cases of depression might contribute to higher participation in the labour market. In light of the scarce evidence available, the present study is one of the first to suggest that depressive symptoms are a significant predictor of total work nonparticipation among both males and females. With regard to LTSA, UE and ER, our findings support the results of the few previous studies. Although administrative data on the diagnoses behind cases of sickness absence and disability pension are largely available, it is not known to what extent these diagnoses reflect a role of depressive symptoms in relation to sickness absence and disability pension in working populations. In a similar vein, it is not known what role depressive symptoms play relative to other types of pensions and to UE. To understand the possible impact of depressive symptoms, approaches quantifying associations with subsequent work nonparticipation are needed that go beyond the calculation of rate ratios (or hazard ratios/odds ratios) for the simple occurrence of events. In the present study, the adoption of two-part models represents an advancement as it considers not only the incidence or prevalence of nonparticipation events but also their duration. This approach could be relevant for the assessment of the economic consequences of depressive symptoms on participation in working life. When examining such consequences, it is important to take all types of work nonparticipation into account to reduce effect underestimation. --- Data Availability A scientific use file (SUF) containing both wave 1 and wave 2 of the cohort is available at the Research Data Centre of the Federal Institute of Occupational Safety and Health. Acknowledgements: The longitudinal Study on Mental Health at Work (S-MGA) was conducted in collaboration with the Institute for Employment Research (IAB). The S-MGA was based on samples from statistics of the Federal Employment Agency (BA), which have been merged with Integrated Employment Biographies by the IAB. A scientific use file of the first wave and second wave data is available at the BAuA (DOI: https://doi. org/10.48697/smga.w1w2.suf ). --- Author contributions Conceptualization, U.R., N.K. and H.B.; methodology, U.R., N.K. and H. B; software, U.R. and N.K.; validation, U.R.; formal analysis, U.R., N.K., and D.P.; resources, U.R. and H.B.; data curation, U.R. and D.P.; writing/original draft preparation, U. R, H.B. P.M.C, D.P. and N.K. project administration, U.R.; funding acquisition, U.R. All authors have read and agreed to the published version of the manuscript. --- Declarations Ethical approval and consent to participate Institutional Review Board Statement: This study was conducted according to the guidelines of the Declaration of Helsinki and was approved by the Ethics Committee of the Federal Institute for Occupational Safety and Health (BAuA; approval number 006_2016_Müller, 6 March 2016). --- Informed consent Informed consent was obtained from all subjects involved in the study. All employees in the sample were contacted by mail, and the interviews at baseline were only conducted after each respondent gave their informed oral consent. Written consent was given for the willingness to participate at follow-up. --- Consent for publication Not applicable. --- Competing interests The authors declare no competing interests. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background We examined the association of depressive symptoms with subsequent events -and duration thereof -of work nonparticipation (long-term sickness absence, unemployment and early retirement).We employed a 5-year cohort from the Study on Mental Health at Work (S-MGA), based on a random sample of employees subject to social contributions aged 31-60 years in 2012 (N = 2413). Depressive symptoms were assessed at baseline through questionnaires, while work nonparticipation was recorded in follow-up interviews. Associations of depressive symptoms with subsequent events of work nonparticipation were examined in two-part models, with events analysed by logistic regressions and their duration by generalized linear models.Medium to severe depressive symptoms were associated with events of work nonparticipation (males Odds Ratio [OR] = 3.22; 95% CI = 1.90-5.45; females OR = 1.92; 95% CI = 1.29-2.87), especially with events of longterm sickness absence in both genders and events of unemployment in males. Mild depressive symptoms were also associated with events of work nonparticipation (males OR = 1.59; 95% CI = 1.19-2.11; females OR = 1.42; 95% CI = 1.10-1.84). Among those experiencing one or more events, the duration of total work nonparticipation was twice as high among males [Exp(β) = 2.06; 95% CI = 1.53-2.78] and about one third higher [Exp(β) = 1.38; 95% CI = 1.05-1.83] among females with medium to severe depressive symptoms.The present study focuses on both events and duration of work nonparticipation, which are both critical for examining societal consequences of depressive symptoms. It is key to regard also mild depressive symptoms as a possible risk factor and to include different types of work nonparticipation.

Introduction Men who have sex with men (MSM) remain the group most affected by HIV in the United States (U.S.), accounting for approximately 63% of all new HIV infections and 52% of all persons living with HIV (Centers for Disease Control and Prevention [CDC], 2012a, 2012b). Among MSM living with HIV it has been estimated that, depending on their race/ethnicity, 19%-26% are unaware of their HIV infection (Chen et al., 2012). The CDC has routinely suggested at least annual HIV screening for sexually active MSM (Branson et al., 2006); however, results from the National HIV Behavioral Surveillance System in 21 cities indicated that 53% of the 680 MSM who tested HIV-seropositive as a part of the study had previously tested more than 12 months prior to diagnosis or had never tested before (CDC, 2011). In the past few years there has been a growing interest in the "treatment cascade" and implications for HIV prevention related to attrition at each step (Mugavero, Amico, Horn, & Thompson, 2013). The "treatment cascade" starts with estimates of prevalent HIV cases, with successive bars illustrating the subsequent steps of HIV diagnosis, linkages to care, retention in care, antiretroviral therapy receipt, and plasma viral suppression. People who are HIV-seropositive but remain undiagnosed are unable to receive the care and treatment they need to slow disease progression and prevent transmission. Early diagnosis of HIV has significant personal and public health benefits. Diagnosis early in disease progression increases survival, improves health outcomes, and decreases health care costs (Chadborn, Delpech, Sabin, Sinka, & Evans, 2006;Hogg et al., 2001;Krentz, Auld, & Gill, 2004;Sterling, Chaisson, Keruly, & Moore, 2003). Delayed diagnosis among MSM has additional implications for disease progression and transmission prevention. MSM who are unaware of their infection are more likely to engage in sexual risk-taking (Marks, Crepaz, Senterfitt, & Janssen, 2005), placing others at risk for acquiring HIV, themselves at risk for contracting other sexually transmitted infections (STIs),weakening their already compromised immune systems (Wiley et al., 2000), and increasing the potential for superinfection with multiple strains of HIV (Sidat et al., 2008). Additionally, a recent study evaluating the primary drivers of the continued HIV epidemic among MSM estimated that 24-31% of new HIV infections among MSM can be attributed to having sex with partners who are unaware of their HIV infection (Goodreau et al., 2012). Increasing attention in the HIV risk-taking literature is being paid to establishing and testing HIV risk models that address not only sociodemographic risk factors, but also personal and contextual factors (Traube, Holloway, & Smith, 2011;Zea, Reisen, Poppen, & Bianchi, 2009). While this literature is primarily focused on behavioral risks for HIV acquisition or transmission, the same theoretical considerations can be taken into account when conceptualizing reasons for delayed HIV diagnosis. A theoretical approach that addresses personal and contextual factors that has received growing support in the HIV-risk literature is Social Action Theory (Ewart, 1991). Applying Social Action Theory to delayed HIV testing, one could propose that self-regulatory action (e.g., getting tested for HIV) is influenced not only by the characteristics of the person (e.g., self-efficacy to engage in protective health behaviors), but additionally by the social, physical, and affective context in which the action occurs (e.g., social support, housing, mental health status). Sociodemographic characteristics of people with late (commonly defined as HIV diagnosis within 12 months of AIDS diagnosis) and delayed (commonly defined as HIV diagnosis one year or more after HIV infection) HIV diagnosis often have been assessed. There has been limited research addressing personal and contextual factors that may affect delayed testing (Kozak, Zinski, Leeper, Willig, & Mugavero, 2013;MacKellar et al., 2005;Mukolo, Villegas, Aliyu, & Wallston, 2013;Nelson et al., 2010). In this manuscript we use qualitative and quantitative data from the Seattle Area MSM Study (SAMS) to conduct a preliminary assessment of potential personal and contextual factors that may contribute to delayed HIV diagnosis among MSM. Specifically we hypothesized that participants who reported increased personal and contextual stressors would be more likely to have delayed diagnosis. Stressors were conceptualized as factors that exhibit the potential to induce negative changes in the individual's psychological, physiological and immunological equilibrium (Paterson & Neufeld, 1989). To enrich our understanding of the relation between the level of stressors and delayed status we compare overall patterns of stressors between delayed versus non-delayed testers as well as provide prototypical case examples. (Thiede et al., 2009). Briefly, potential participants were passively referred to the study after testing for HIV at the PHSKC sexually transmitted diseases (STD) and HIV/AIDS Program (HAP) clinics, HAP community outreach testing sites, and two university HIV clinics. MSM were eligible if they were age 18 or older, reported sex with men during the preceding 6 months, and were able to complete the interview in English. HIV-seropositive MSM were eligible if recruited within three months of their first HIV-seropositive test result and HIV-seronegative MSM were eligible if recruited within one month of testing. A structured quantitative survey questionnaire was administered to all participants via audio computer-assisted selfinterviewing (ACASI). All HIV-seropositive and half of the HIV-seronegative participants were asked to participate in qualitative semi-structured interviews. --- Method All interviews were audio recorded with concurrent supplementary written notes taken by the interviewer. Within a week of each interview, the interviewer would create a write-up of the interview including their notes. These write-ups would include a combination of descriptions of the substantive content of the discussion along with verbatim transcribed quotes and notes on the context of the quotes. Verbatim quotes were chosen based on their direct relation to the study goals, whereas lengthy, less relevant, or off-topic sections of the interviews were summarized by the interviewer in the write-up. This method has a long history of successful and reliable use within traditional ethnography and is an efficient way to contain write-up costs (Spradley, 1979). To validate their accuracy, write-ups were reviewed by two additional research team members. To establish and contrast factors associated with delayed and non-delayed testing we conducted a content analysis using CDC EZ-Text software (Carey et al., 2008). Two coders developed a codebook designed to identify themes in the interview write-ups (Hruschka et al., 2004). Overall, final intercoder reliability was that 85.4% of the codes had kappa values equal to or greater than 0.8. The codes were used to identify common behavioral patterns and to search the database for the illustrative verbatim quotes presented below from the delayed and non-delayed tester samples. During the study period all persons testing HIV-seropositive at the PHSKC sites were routinely offered the less sensitive (LS) HIV-1 test, employing the serologic testing algorithm to assess recent HIV seroconversion (STARHS), to estimate recent infection in the testing population. All tests were performed by the PHSKC Laboratory (Vironostika-LS EIA; Bio Merieux, Raleigh, NC). The estimated mean time from seroconversion (defined using a standardized optical density cutoff of 1.0) with the Vironostika-LS is 170 days (95% CI, 145-200 days; Kothe et al., 2003). To be consistent with our previous manuscript on delayed diagnosis among this population (Nelson et al., 2010), delayed testers were defined as MSM who seroconverted six months or more before testing HIV-seropositive, based on a positive HIV-1 test combined with a reactive LS HIV-1 test (according to STARHS) or, in the case of a missing LS HIV-1 test, lack of a self-reported HIV-negative test during the past six months. Non-delayed testers were defined as those with infections acquired within the preceding six months, based on a positive HIV-1 test combined with a non-reactive LS HIV-1 test (according to STARHS) or self-reported HIV-seronegative test during the past six months. Among participants in whom we relied on self-reported last HIV-seronegative test, attempts were made to verify the date of last test by contacting the participant's medical provider or through PHSKC HIV testing records. Through this method we were able to verify the date of last HIV-seronegative result for three of the non-delayed testers. In addition, six of the non-delayed testers were verified to be recent infections due to their diagnosis with a syndrome of primary infection at an HIV clinic that specialized in diagnosis and treatment of new infections in addition to their self-reported prior HIV-seronegative tests. Two participants had reactive LS HIV-1 results while having a self-reported HIVseronegative test in the past six months. In these cases recency of infection was decided by the study investigators based on information the client provided in the quantitative and qualitative interviews about their testing history and usual frequency of testing. Given these participants' reported frequency of testing and the amount of details, including dates and locations, provided about their last HIV-seronegative test result, these two participants were deemed to be non-delayed testers. Overall, quantitative and qualitative analyses were conducted on the 75 of the 77 participants who tested HIV-seropositive and had coded qualitative interview data, including 38 delayed and 37 non-delayed testers. This study was approved by the Institutional Review Boards of the University of Washington and the CDC. All participants provided informed consent. --- Analysis To create a description of our sample, chi-square and Fisher's Exact Tests were used to compare delayed and non-delayed testers with respect to sociodemographic variables. Sociodemographic variables included: recruitment site (HAP clinic, HAP outreach, STD Clinic, and other sites), age (18-29, 30-39, 40+), race/ethnicity (white, black/African American, Hispanic, other), education (high school or less, more than high school), sexual identity (gay, bisexual), extent they had revealed their sexual practices ("out" to >50% of the people they knew about having about male-male sex, "out" to <unk>=50% of the people they knew about having male-male sex), and health insurance (yes/no). To establish variables that would likely increase the level of stressors in a participant's life, we analyzed two questions from the semi-structured interview: 1) "Before you found out your HIV results, had anything big happened in your life in the last year? For example, did you move, lose or start a job, did a relationship end or begin, etc." and 2) "Do you have any history of depression or other mental health issues?" All participants were asked these questions with additional specific probes to further engage the participant as necessary. For example, for the question "Before you found out your HIV results, had anything big happened in your life in the last year?" the probes included: 1) "Was there anything from before a year ago that had a bigger impact for you?"; 2) "What happened?"; 3) "Where did it happen?"; 4) "When did it happen?"; 5) "How did it happen?"; 6) "Who was there?"; 7) "How did you deal with this?"; 8) "What helped you cope?"; and 9) "How did things work out?" Four main variables in the qualitative data were hypothesized to increase the level of stressors in a participant's life. These included any experiences of violence (yes/no), relationship issues (yes/no), or substance use issues (yes/no) in the past 12 months and a self-reported history of hospitalization for mental illness (yes/no). Participants were coded as having experienced violence if they reported: 1) doing something violent, acting out, or being malicious; 2) being physically abused or hurt; 3) being sexually abused, molested, or raped; or 4) being verbally or emotionally abused or hurt in the past 12 months. Relationship issues were defined as participants who self-reported experiencing significant events in their romantic, family, or friendship relationships (e.g. starting a new relationship, ending a relationship, relationship conflict) in the past 12 months. Participants who self-reported increases in illicit drug or alcohol use, attempting to become sober, overdosing, and/or entering detoxification treatment or other substance use related programs in the past 12 months were coded as having substance use issues. To supplement the qualitative data we additionally identified behavioral variables in the quantitative dataset that we hypothesized would likely increase stress in participant's lives. These included current homelessness (yes/no) and history of incarceration (yes/no). Homelessness was defined as self-reporting being homeless or having a non-permanent residence. History of incarceration was defined as self-reporting ever having spent a night in jail. A count variable was created to represent the overall level of stressors experienced in a participant life, with a 0 corresponding to not having experienced any of the hypothesized stressors and a 6 corresponding to having experienced all of the hypothesized stressors. Participants were considered to have encountered significant stressors in their lives if they experienced 2 or more of the proposed stressors. Fisher's Exact Tests were used to compare delayed and non-delayed testers with respect to the individual stressors and the overall stressor variable (0-1, 2+). Overall patterns, case examples, and illustrative quotes were extracted from the qualitative data to enrich our understanding of the relation between the level of stressors and delayed status. We established four case examples: a HIV-seropositive participant who was delayed and reported 2 or more stressors, a HIV-seropositive participant who was non-delayed and reported 2 or more stressors, a HIV-seropositive participant who was delayed and reported 0-1 stressor, and a HIV-seropositive participant who was non-delayed and reported 0-1 stressor. --- Results --- Sample Among the 77 HIV-positive MSM enrolled in SAMS, 75 had coded qualitative interview data, including 38 (51%) delayed and 37 (49%) non-delayed testers. As reported previously (Nelson et al., 2010), delayed testers were more likely to be older (18-29 yrs old: 21% vs. 30%, 30-39 yrs old: 29% vs. 54%, and 40+ yrs old: 50% vs. 16%, p<unk>.01), black/African American (26% vs. 3%, p<unk>.01), and "out" to 50% or less of the people they knew about having male-male sex (45% vs. 14%, p<unk>.01) compared to non-delayed testers (Table 1). --- Proposed stressors and their relation with delayed diagnosis An analysis of the proposed stressors in the qualitative and quantitative data revealed that delayed testers were no more likely to report having experienced violence (26% vs. 19%, p = 0.58), relationship issues (53% vs. 51%, p = 1.0), or substance use issues (34% vs. 30%, p = 0.81) in the past twelve months; a history of hospitalization for mental illness (21% vs. 14%, p = 0.54); or a history of incarceration (50% vs. 30%, p = 0.10) compared to nondelayed testers (Table 2). Delayed testers were more likely to be homeless (32% vs. 5%, p = 0.01). Interestingly, while the majority of the proposed stressors were not significantly different between delayed and non-delayed testers, delayed testers were significantly more likely to have experienced two or more of the stressors (68% vs. 38%, p = 0.01) compared to non-delayed testers. Upon further exploration of the narratives provided by the participants multiple patterns arose. Specifically, delayed testers with two or more life stressors often described a long history of substance use and untreated or only sporadically treated mental illness. Further they reported feeling that they felt a lack of agency of control over their own behaviors. Lastly, they reported a lack of social support and accurate appraisal of their HIV risk. In contrast, non-delayed testers with two or more life stressors described similar issues with substance use, mental health, and mental health treatment, though, non-delayed testers tended to report a sense of control and agency over their own behaviors as well as an accurate appraisal of their HIV risk. Delayed testers with fewer than two life stressors often described no or few substance use issues; overall good mental health and/or a stable mental health treatment history; having a sense of agency over their own behaviors; and having a steady social support system. Similar to delayed testers with two or more life stressors, delayed testers with fewer than two life stressors also appeared be to have an inaccurate appraisal of their own HIV risk, often due to a false sense of security in the context of a long-term relationship. Non-delayed testers with less than two life stressors on the other hand, also reported having little-to-no substance use issues, stable mental health or mental health treatment history, a stable social support system, and a sense of agency over their own behaviors, but, in contrast to the delayed testers with less than two life stressors, they tended to be accurately aware of their own HIV risk. --- Case Examples One HIV-seropositive delayed tester who experienced 2 or more stressors in his life was black/African American, homeless, 30-39 year old, had a high school diploma or lower level of education, and did not have health insurance. He described himself as experiencing a lifetime of suffering from substantial alcohol use and mental health issues, including being hospitalized on multiple occasions. Further he reported having emotional and financial problems related to his alcohol use. He described spending all of his money on alcohol and being unable to afford things like housing or food. Further, he said that his drinking isolates him from his family as they do not want to be around him when he is drinking because he gets "obnoxious and smart" when he drinks. In the year prior to testing he reported moving to the Seattle area to seek work, but remained unemployed. He described himself as sad about being away from family and overall feeling socially isolated and lonely; "I don't have too many friends." He reported that he has sex with men not because he wants sex, but instead as a means of having someone to talk to and be physically close with. He described meeting most of his sexual partners in bars or public parks and that he is incapable of meeting, conversing, or having sex with people without being drunk. He reported feeling used by men and that the men he has sex with are only having sex with him to get something from him (e.g., place to stay, money). He stated that they "are not my friends." A HIV-seropositive non-delayed tester who reported 2 or more stressors was white, homeless, 30-39 years old, who had a high school diploma or lower level of education, and no health insurance. He reported a long history of mental health issues that he had attempted, without much success, to get treated through the public health system. He reported significant drug use through his lifetime, including pot, speed, alcohol, and heroin and that most people would describe him as a -"junkie, thief, liar, hypocrite, careless, liar, scared, immature..." He reported testing because his primary partner had recently tested HIV-seropositive. He reported feeling that it is "the destiny of the gay population to go out with AIDS" and that gay people deserve AIDS because "it is unnatural, unhealthy for the gay community to identify itself with their sexual practices..." He further reported that in the past he had reduced his risk of contracting HIV by only letting "love affair" partners ejaculate inside him. He didn't ask his sex partners about their HIV statuses, and said "I don't want to know." He feels that even if he asked them, they could be lying, so he assumes all his partners are positive. He was a prostitute for a long time, and felt "invincible" from HIV because he did not pick it up while hustling. However, at this point in his life, he felt very "used up" and described having to "work to make sex a positive experience." A HIV-seropositive delayed tester who reported 0-1 stressors was a Hispanic, 40+ year old, who lived with a spouse/partner, is a college graduate, and has health insurance. In the past, he reported being treated for depression after breaking up with a partner of 11 years. He stopped seeking services (therapy & medication) once he recovered, and had not had problems with depression since then. He described himself as "sincere, honest, warm, generous..." He described liking monogamous, long term relationships, having four 3+ year relationships/sex partners since he came out as gay. He reported that he did not like to do drugs or party. He noted that he and his most recent partner did not have sex until they had been dating for three months. After two years, they stopped using condoms because his partner told him, the "last time I tested, I tested negative" before they started having sex and he felt "like he was my soul mate -that this was pretty much it for me." He reported not feeling at risk because of this. He did not identify anything stressful happening to him in the year prior to testing HIV-seropositive. A HIV-seropositive non-delayed tester who reported 0-1 stressors was white, under 25 years old, living with roommates or friends, and was a college graduate with no health insurance. He reported ongoing medication management for an anxiety disorder and a history of depression that resolved three years ago. He related his depression to the stress experienced after coming out. He said that all the mental energy it was taking to appear straight was suddenly available for other things -"I came out -and I think it was like the first time that I was using my whole brain! So it was a little overwhelming." He reported routinely testing for HIV and that he met the man he believes infected him at a bathhouse. He described not normally meeting men in that kind of venue, but was feeling curious and it was available. He was about to leave the area after graduating college and wanted to experience that element of gay culture. "Why am I faced with this issue and it doesn't seem to be there for anybody else. The immediacy of if you're horny and want to have sex, if you're gay -it's so easy... for straight guys it's just totally not." He described that the existence of bathhouses is wrong and dirty, and he has to be tipsy to go to them. He did not identify anything stressful happening to him in the year prior to testing HIVseropositive. --- Discussion HIV testing is one of the primary HIV prevention strategies supported by the U.S. government. With the accessibility of HIV testing as well as the significant personal benefits of timely HIV diagnosis, it is remarkable that a substantial proportion of MSM delay HIV testing and remain unaware of their HIV infection (Chen et al., 2012). While previous research in this area has attempted to understand the sociodemographic factors that are related to delayed HIV diagnosis, limited work has attended to personal and contextual factors (Kozak et al., 2013;MacKellar et al., 2005;Mukolo et al., 2013;Nelson et al., 2010). Our study sought to provide a preliminary view of the complicated interactions between personal and contextual factors that may lead to delayed testing among MSM. In line with Social Action Theory, in our sample, getting tested for HIV appears to be influenced by a confluence of personal characteristics (e.g., self-efficacy to engage in protective health behaviors), as well as the social, physical, and affective context in which the testing occurs (e.g., social support, housing, mental health status). Specifically, we found that the majority of the personal and contextual factors we hypothesized would increase the level of stressors in a participant's life may not be individually associated with delayed HIV diagnosis. However, participants who indicated that they experienced two or more stressors in the past year were significantly more likely to have delayed testing. Additionally, in our further exploration of the participants' narratives we identified some prominent patterns related to the stressor-delayed diagnosis relation. Specifically, it appears that there may be a few factors that differentiate those who are able to cope with multiple stressors and continue to engage in timely health care seeking behaviors versus those who are not. Our data indicates that men with stable social support and mental health as well as self-efficacy to engage in protective health behaviors may be more equipped to cope with multiple stressors and continue to engage in timely health care seeking behaviors. Whereas, men who are experiencing multiple stressors but are lacking in these supportive factors may not have the coping skills necessary to engage in timely health care seeking behaviors. Interventions attempting to address delayed diagnosis among MSM should specifically attempt to target social support, mental health, and self-efficacy to engage in protective health behaviors as well as integrate techniques to address coping skill deficits. Additionally, it is possible that MSM who are experiencing multiple life stressors without the scaffolding of social support, stable mental health, and self-efficacy to engage in protective health behaviors are unable to prioritize HIV testing over more salient and pressing issues. It is likely that MSM who are experiencing multiple stressors, such as being homeless and experiencing violence, feel a need to address their physiological and safety concerns before they are able to prioritize HIV testing. If this is true, it indicates the further need for structural interventions addressing these issues in addition to interventions to increase social support, mental health treatment, and self-efficacy to engage in protective health behaviors. Lastly, it appears that recognition of HIV risk is an important factor in delayed diagnosis. Delayed testers, regardless of level of stressors, predominantly indicated that they were unaware of their HIV risk. This result is in line with previous research which found that the majority of MSM either with late diagnosis or undiagnosed HIV infection perceived themselves to be at low risk for acquiring the virus (MacKellar et al., 2005;Schwarcz et al., 2011). While obsessing about HIV risk is likely to be maladaptive, averting attention from it may lead to delayed testing and increased risk-taking. As we found in our previous work (Nelson et al., 2010), many of the delayed testers reported that they did not frequent MSMspecific venues or identify strongly with the MSM community. As such, effective messaging and interventions hoping to increase recognition of HIV risk among MSM who delay testing may need to target men generally instead of MSM specifically and be distributed more widely to locations that are non-MSM specific. To our knowledge there has only been one previous study exploring personal and contextual factors related to delaying testing for HIV (Schwarcz et al., 2011). In a recent qualitative and quantitative study of 41 persons who developed AIDS within 12 months of their HIV diagnosis (i.e., late testers), Schwarcz et al. (2011) found that the barriers to testing were multiple and often interwoven. Barriers in this study included fear of receiving an HIV diagnosis, concerns about ability to pay for treatment, inaccurate knowledge about treatment, inaccurate assessment of HIV risk, and the presence of mental health issues. Because late testers are a subset of delayed testers, it is likely that many of the same personal and contextual factors associated with late testing may also be associated with delayed testing; these patterns were partially supported by our previous findings (Nelson et al., 2010) and in this current investigation. As with any study there are limitations that should be kept in mind when considering these results. First, due to our limited sample size, and thus limited power, it is possible that some of the individual characteristics that were assessed (e.g., history of incarceration, history of hospitalization for mental illness) may in fact be individually associated with delayed diagnosis. As such, we consider this a preliminary assessment of these factors and encourage further research with a larger sample size to fully explore the influence of these characteristics on delayed diagnosis. Second, we recognize that our data is now approximately ten years old and awareness of HIV risk as well as HIV testing patterns may have changed since we collected our data. While this is the case, we feel that these personal and contextual factors are likely to continue to have an influence the timeliness of HIV testing and should be further explored. Third, the use of self-reported behaviors and inperson qualitative interviews may be prone to social desirability and recall bias. Fourth, due to the demographic make-up and the recruitment locations (Thiede et al., 2009), the study population may not be representative of all MSM in the broader community who are recently diagnosed with HIV. Finally, there are potential limitations in our use of a crosssectional design and the associated classification of delayed and non-delayed testers. Given the window period for the Vironostika-LS HIV-1 test using STARHS (145-200 days) it is possible that while a person was reactive on the LS HIV-1 test they in fact seroconverted within the past six months (Kothe et al., 2003). Additionally, as not all participants were evaluated using STARHS, there were some cases where we relied on self-reported HIV testing history to establish whether a person seroconverted within six months of testing or not. Thus, it is also possible that while a person did not test within the past six months, he, in fact, seroconverted within the past six months. While this may be the case, as reported previously (Nelson et al., 2010), the majority of participants classified as delayed testers identified an illness associated with HIV or an illness that persisted much longer than what would normally be expected in an immunocompetent person as their reason for testing, in addition to the majority having not tested within the past two years, it seems likely that most delayed testers seroconverted longer than six months prior to their HIV-seropositive test. Overall our findings indicate that personal and contextual factors likely work in concert to influence delayed HIV diagnosis among MSM. Specifically, MSM who experience multiple life stressors, whether they are personal or contextual, have an increased likelihood of delaying HIV diagnosis. Additionally, factors such as social support, mental health, and selfefficacy to engage in protective health behaviors as well as an accurate recognition of HIV risk appear to be of particular importance, making them ideal targets for intervention. As early HIV diagnosis is important on both a personal and community level, interventions targeting the enhancement of coping skills, self-efficacy to engage in protective health behaviors, social support, and mental health treatment as well as structural interventions targeting physiological and safety concerns will likely help MSM handle their life stressors in a more effective manner and engage in more positive health behaviors, including seeking HIV testing in a more timely manner. Note. Violence = 1) doing something violent, acting out, or being malicious; 2) being physically abused or hurt; 3) being sexually abused, molested, or raped; or 4) being verbally or emotionally abused or hurt; Relationship issues = significant events in their romantic, family, or friendship relationships; Substance use issues = illicit drug or alcohol use, attempting to become sober, overdosing, and/or entering detoxification treatment or other substance use related programs; Homelessness = being homeless or having a non-permanent residence; History of incarceration = having spent a night in jail.

Delayed HIV diagnosis among men who have sex with men (MSM) in the United States continues to be a significant personal and public health issue. Using qualitative and quantitative data from 75 recently tested, HIV-seropositive MSM (38 delayed and 37 non-delayed testers) we sought to further elucidate potential personal and contextual factors that may contribute to delayed HIV diagnosis among MSM. Our findings indicate MSM who experience multiple life stressors, whether personal or contextual, have an increased likelihood of delaying HIV diagnosis. Further, MSM experiencing multiple life stressors without the scaffolding of social support, stable mental health, and self-efficacy to engage in protective health behaviors may be particularly vulnerable to delaying diagnosis. Interventions targeting these factors as well as structural interventions targeting physiological and safety concerns are needed to help MSM handle their life stressors more effectively and seek HIV testing in a timelier manner.

Introduction Participatory research denotes the engagement and meaningful involvement of the community of interest in multiple stages of investigation, from design to data collection, analysis, and publication. In the 1970s, participatory action research (PAR), influenced by the work of Brazilian author Freire (1987), emerged in the social sciences to challenge the neutrality of science and address power asymmetries between academic and popular knowledge (Borda, 2008). PAR proposes a North-South encounter based on solidarity and mutual enrichment paying special attention to the colonial legacy of oppressed countries. According to de Sousa Santos (2015), an understanding of the world beyond the western understanding of the world exists; we need cognitive justice to have global social justice; and emancipation takes shape in diverse ways outside of Western theory (de Sousa Santos, 2015). As such, this type of research favors subject to subject interactions toward transformation, rather than subject to object, thus reshaping epistemological notions that underpin traditional research (Fals-Borda, 1987). Mental health research in the Global North incorporated principles of PAR to study problems with communities (Kidd et al., 2018), and a discrete body of knowledge led by people who identified as consumers, survivors, and ex-patients of psychiatry has developed (e.g., survivor research and mad studies) (Faulkner, 2017). The Alma Ata was the first international declaration to explicitly state that individuals have the right to participate in shaping healthcare (WHO, 1978). Since the 1980s, public and patient involvement (PPI) or community and public engagement, grown out of civil rights movements, governmental initiatives, and disability rights movements (Tomes, 2006;Sweeney et al., 2009), further consolidated the importance of participation in shaping public policy and in research (Hickey et al., 2022). This tradition continued to expand in high-income countries (HICs), and acceptance of the principles of participatory research by mainstream science is growing (Pearce, 2021). Examples of the institutionalization of PPI in HICs are the National Institute of Health Research funded INVOLVE in the United Kingdom (started in 1996), the Patient-Centered Outcomes Research Institute (PCORI) in the United States, and the 2009 International Collaboration for Participatory Health Research with all but one steering committee member coming from the Global North (Beresford and Russo, 2020). Participatory research in mental health counters the history of a field defined by the politics of exclusion (Mills, 2014). Unlike other medical specialties, psychiatry's relationship to social order has shaped discourses (e.g., notions of the degenerative course of mental illness, mental illness as a moral problem, and mental illness as deficit), and interventions (e.g., long-term internment and involuntary hospitalizations) that resulted in the systematic exclusion of individuals from society (Foucault, 1999). While a shift from asylums as the privileged locus of "treatment" happened globally, and deinstitutionalization became a priority, the issue of social exclusion has not been resolved, and even as co-production and participation develop in research, power asymmetries remain (Rose and Kalathil, 2019). Research participation in low-and middle-income countries is further complicated by the enduring history of colonial legacies (Bulhan, 2015) and the geopolitics of knowledge (Naidu, 2021) that privileges Euro-American epistemologies (Mignolo, 2005), which materializes in science in multiple ways. These have multiple consequences: globally, more than 98% of all funding streams for mental health are awarded by HICs (Woelbert et al., 2020) who consequently occupy a privileged position with respect to knowledge production (Abimbola, 2019); representation of individuals who are white, male, and HIC-based working at academic places of power (e.g., editorial boards and universities) is disproportionate (Naidu, 2021). In addition, interventions developed in the Global North have been systematically adapted and implemented in the Global South, for example, the World Health Organization's mhGAP (Timimi, 2011). An equally consequential form of power and exclusion lies in the epistemological dominance of HICs, who determine the methods and knowledge that count, and continue to exclude, silence, and oppress forms of being and knowing originating in the Global South (Alejandro Leal, 2007;Bulhan, 2015;Bhakuni and Abimbola, 2021;Naidu, 2021). Psychiatry's history of exclusion and marginalization combined with the epistemic violence that results from the geopolitics of knowledge that shape the North-South relationship places participatory research in mental health in the Global South at the intersection of multiple oppressions. The psychiatric reform movement in Brazil took place in the context of broader societal changes toward democratization. Starting in the 1970s, mental health workers in the country organized to denounce the abuse and inefficiency of psychiatric hospitals to treat and support the recovery of people with severe mental health problems (Amarante, 1998). Inspired by Basaglia's Democratic Psychiatry and the experiences of deinstitutionalization in Italy, the anti-asylum movement grew side by side with Brazil's universal public health system, both informed by a strong critique of positivist and biomedical epistemologies as insufficient to address social problems (Yasui, 2010;Amarante, 2015). This paradigmatic shift from asylums to the psychosocial care system was enshrined in law in 2001 (Law 10.216). Service user participation is a key feature of the Brazilian public health system and built into the principles of the psychiatric reform movement as well as the public healthcare system. Despite numerous successful experiences of service user involvement and leadership in mental health policymaking, service delivery, and advocacy (Vasconcelos, 2009), effective and consistent participation remains aspirational. Participation in research, however, is not as well established. More than 20 years since the shift in how mental health services are organized in Brazil has been enshrined in law, evaluation of mental health services using participatory methods remains scarce (Ricci et al., 2020). This historical context makes it so that participation in policymaking, advocacy, and research are not understood separately. The state of the art of mental health research in the country is unknown. This knowledge gap is problematic locally and globally. Locally, participatory initiatives remain isolated in the context of specific projects and a national agenda for the advancement of participation of service users and people with lived experience of mental health problems would benefit from this scientific knowledge base. Globally, researchers remain unaware of the wealth of knowledge produced in Brazil. Thus, our study sought to review the empirical participatory literature in mental health in Brazil, identify common themes, and synthesize the results. --- Objective Our scoping review's objective was to chart and analyze the empirical participatory mental health research literature in Brazil. Our review focused on research studies and included gray literature. Specific objectives were to describe how participation is conceptualized in mental health research in Brazil, to identify key concepts associated with participatory research, and to identify the main obstacles to participatory research in mental health in the country. --- Methods Our team included academics in Brazil, Chile, the United States, and the United Kingdom, and people with lived experience of mental health challenges. Our review followed the Joanna Briggs Institutes' guidelines (Peters et al., 2017) and PRISMA extension guidelines for scoping reviews (Tricco et al., 2018). --- Eligibility criteria Eligibility included studies that used participatory research methods, broadly defined as research in which service users' and family members' roles in the study went beyond those of research subjects (i.e., an individual who provides information or data to help answer a research question). We defined participatory research procedures broadly and included studies that employed member-checking, consultations, co-production, data validation procedures, and stakeholder consultation groups. We included empirical studies conducted in Brazil and published in peerreviewed journals in Portuguese, Spanish, English, or French (study team's languages). We included the gray literature as well. We did not specify dates. Studies using quantitative, qualitative, or mixed methods were eligible. Studies that included participants of any age, sex, gender, ethnicity, race, or class were included. Studies involved service users and/or family members with any mental health diagnosis but excluded those with a primary diagnosis of a physical health problem (e.g., epilepsy and dementia) or substance use exclusively. We excluded studies that claimed to be participatory but provided no evidence of participation in the methods or results section of the paper. We excluded studies focusing on providers only but kept those that included providers if families or service users were included as well. --- Information sources Our initial exploration of the topic revealed a series of challenges to traditional search strategies. Key words and vocabulary were inconsistent, metadata were missing, and titles were not available in databases (e.g., Web of Science and Scopus). To address these challenges, our team developed a multipronged strategy that combined bidirectional citation tracking (Hinde and Spackman, 2015) and a targeted search strategy to a variety of databases to identify relevant studies. We conducted two bidirectional searches and one additional search of targeted databases. The research team, through their knowledge about this literature, Google Scholar searches, and consultations with experts, first identified 10 relevant studies (known as "pearls"), and the medical librarian used these pearls to identify articles through a systematic search of their cited and citing articles using citationchaser, SciELO, Scopus, Web of Science, and Google Scholar. Using the included studies, we identified relevant terms using the Systematic Review Accelerator WordFreq tool (Clark et al., 2020) and developed a targeted search strategy. The librarian searched the following databases: MEDLINE, Embase, PsycInfo, Global Health, LILACS, Web of Science, Scopus, SciELO, BDTD, and the PBiPortal de BuscaIntegrada. We limited search results to English and Portuguese titles because the previous step did not yield relevant results in French and Spanish. Our team translated the search strategy to Portuguese between the following databases to find published and unpublished (i.e., gray) literature: PubMed, MEDLINE (Ovid), Embase (Ovid), PsycInfo (Ovid), Web of Science Core Collection, Scopus, and Scielo.br. We pooled results in EndNote, removed duplicates, and uploaded to Covidence. We identified relevant theses and dissertations using the National Thesis Database BDTD and relevant books using the Universidade de S<unk>o Paulo library catalog. Finally, our team consulted experts in the field, charted the main publication venues outside mainstream academic databases (ABRASME, APRAPSO, and ABRASCO for conference proceedings), and handsearched key journals (e.g., Revista de Sa<unk>de Coletiva). --- Selection of sources of evidence Two independent reviewers screened studies' titles, abstracts, and full texts using screening checklists that were pilot-tested and adjusted using the first 100 articles. Decision trees helped resolve ambiguous situations during the screening process. Ultimately, the study team decided to exclude the gray literature because most did not present primary studies. Many studies did not describe the methods making it difficult to assess if they were primary studies or not. Relevant theses and dissertations that were empirical research generally had an associated peer-reviewed publication, which we included. --- Data charting process Our team iteratively developed a data charting form using Covidence and Excel and used it to extract relevant information (Tables 1 and2). The first author and a member of the study team extracted the data independently and resolved conflicts together. We consulted a third member of the study team when consensus could not be reached. --- Data items and synthesis of results Synthesis of results was iterative. We extracted data items that were relevant to the objectives of our review first in Covidence and then in Excel, including: article information (i.e., title, authors, year of publication, and aims); demographic information (i.e., age, sex, gender, ethnicity, race, and socioeconomic status); clinical characteristics (i.e., mental health problems) of the sample; whether participants were service users, family members, or providers; the study setting (i.e., community mental health center, primary care, and university); the methodological and analytical approaches; and definitions of the participatory elements in the study (e.g., member checking, designing research questions, and co-production). We used Atlas.ti to free-code the articles, and finally inductively developed a set of categories by grouping and organizing the codes. --- Results We identified 1,437 references through the search strategy. After removing 814 duplicates, we screened the titles and abstracts of 974 references. We assessed 536 full text studies for eligibility and excluded 516 for several reasons (Figure 1), leaving final pool of 20 studies to be charted and synthesized. --- Characteristics of sources of evidence Study publication dates ranged from 2009 to 2021. Most studies were carried out in the South and Southeast areas of Brazil (N = 13) and published in Portuguese (N = 19). Sample sizes ranged from 7 to 420, with a median number of 15 participants. Most studies were conducted at Community Mental Health Centers (N = 14) and employed qualitative methods (N = 18). Studies used a variety of data analysis approaches, with hermeneutic analysis (N = 7) being the most common. Most studies included participants diagnosed with serious mental illness, psychosis, or both (N = 18). None of the studies reported full demographic characteristics (i.e., sex, age, and race/ethnicity). Nine studies reported on a multicentric study in partnership with a Canadian university that translated and implemented a medication management guide in community mental health centers in several regions of Brazil (Jorge et al., 2012;Onocko Campos et al., 2012;Emerich et al., 2014;Silveira et al., 2014;Gonçalves and Campos, 2017;Chassot and da Silva, 2018;Senna and Azambuja, 2019;Palombini et al., 2020;Passos et al., 2020). Seven studies reported using principles of Guba and Licoln's fourth-generation evaluation, a constructivist methodfocused on negotiationthat engages stakeholder groups across all phases in iterative processes to reach consensus (Guba andLincoln, 1989, 2001;Kantorski et al., 2009;Onocko Campos et al., 2009;Jorge et al., 2012;Emerich et al., 2014;Moreira and Onocko-Campos, 2017;Alves et al., 2018;Palombini et al., 2020). Group validation of results was the most common participation strategy (Kantorski et al., 2009;Onocko Campos et al., 2009;Jorge et al., 2012;Emerich et al., 2014;Gonçalves and Campos, 2017;Alves et al., 2018;Massa and Moreira, 2019;Palmeira et al., 2021). Only two studies explicitly stated that participants were involved in all stages of research (Chassot and da Silva, 2018;Vaz et al., 2019). Participants were not co-authors in any of the included studies, nor was there mention of authors' lived experience in any of the studies we included. Descriptions of the value and role of participation in research varied. Authors acknowledged the importance of community participation in public policy as a means to connect research and action (Lima et al., 2014), noting that participation in research addresses power imbalances in the researcher-subject dyad (Passos et al., 2020). Another study marked diversity, respect and differences, and acknowledgment of lived experiences as legitimate sources of expertise as important reasons for participation (Onocko Campos et al., 2009). One study described participation as a tool to increase political reflection, bring attention to the rights participants may have lost, and increase the relevance of research (Moreira, 2021). Studies that employed fourth-generation evaluation methods highlighted the importance of stakeholder involvement in all stages of research to level power asymmetries in research and increase the relevance of knowledge produced (Jorge et al., 2012;Emerich et al., 2014;Alves et al., 2018;Senna and Azambuja, 2019;Palombini et al., 2020). One example of fourth-generation evaluation included service users, managers, psychiatry residents, and family members to translate, adapt, and test a medication management tool for people with serious mental illness. Their inclusion fostered a sense of agency in the research process, and researchers see themselves as social actors sharing the experience of the world and bringing their Cambridge Prisms: Global Mental Health own subjective experiences (Emerich et al., 2014). One study noted the need to increase participation in research, especially in mental health, given that, historically, service users have been excluded from decision spaces including about their own treatment (Gonçalves and Campos, 2017). One study noted that definitions and operationalization of participation vary greatly and can have different meanings (Moreira and Onocko-Campos, 2017). Cambridge Prisms: Global Mental Health --- Synthesis of results --- Importance of participation In the context of the Brazilian psychiatric reform movement, several studies have considered participation from various dimensions: policy, political, and clinical. From a policy perspective, studies note the shift from the biomedical, hospital-centric model of care to the creation of community-based mental health centers (CAPS) as the main locus of treatment in the public mental health system (Serpa Junior et al., 2014). From a political standpoint, participation in research and in shaping services emblematizes autonomy, citizenship, and the general exercise of civil liberties. Finally, the clinical perspective, more directly related to treatment encounters and service delivery, amplifies the political by connecting suffering with exclusion and marginalization; and treatment with freedom, autonomy, and participation in society. It [action research] is an essentially political way of doing research, which aims at the promotion of citizenship and focuses on the processes of social exclusion. (Moreira and Onocko-Campos, 2017, p. 465) Power and knowledge A prevailing theme among selected articles is that psychiatry has historically oppressed and silenced the people it aims to serve by placing excessive emphasis on scientific and professional knowledge, and by reducing individuals to their diagnoses and symptoms (Jorge et al., 2012;Emerich et al., 2014;Gonçalves and Campos, 2017;Moreira, 2021). Deemed incapable of living in society, subjects are silenced and their tragic experience, frustration, failure and everyday suffering are gradually removed from daily experience and turned into psychopathological categories. (Moreira, 2021(Moreira,, p. 1190, our translation), our translation) Studies point to harmful diagnostic language and treatment practices, rooted in a reductionist biomedical model, that have harmed and violated the rights of people with mental health problems. The separation between knowledge and the experience of madness legitimized Psychiatry's knowledge supremacy and made interventions also an expression of a power-knowledge in the name of treatment. (Moreira, 2021(Moreira,, p. 1190, our translation), our translation) Authors suggest a need to correct power imbalances as critical to advance mental healthcare. These perspectives are strongly rooted in the works of Foucault and Basaglia. --- Autonomy and empowerment Empowerment has historical roots in the struggle for civil rights in Brazil starting in the 1970s. Grounded in the work of Freire (1987) and popular education, this tradition motivated the public health and mental health reform movements to transform traditional forms of power and knowledge and foreground the rights of the historically oppressed (Garcia et al., 2017). In mental health, authors note that empowerment can be paradoxical because the need for special social rights (e.g., benefits, protected work, and free transportation) often clashes with universalist claims of civil rights (e.g., equal rights, social inclusion, and full participation in society) due to the extreme disenfranchisement of populations with intersecting vulnerabilities (e.g., extreme poverty, psychiatric diagnosis, violence, and food insecurity) (Gonçalves and Campos, 2017). In a country in which the precarity of access to social rights for survival is constant, the service user in intense psychic suffering seems to, often, experience a double process of exclusion: to be Brazilian and to be mad. (Emerich et al., 2014, p. 686, our translation) In clinical care, lack of empowerment means not having enough information to make decisions about treatment. This is reinforced by power imbalances favoring professional and academic knowledge (Onocko Campos et al., 2012). Increasing autonomy is an important treatment outcome in the selected studies (Kantorski et al., 2009;Onocko Campos et al., 2009;Jorge et al., 2012;Emerich et al., 2014;Lima et al., 2014;Gonçalves and Campos, 2017;Alves et al., 2018;Chassot and da Silva, 2018;Senna and Azambuja, 2019;Palombini et al., 2020;Passos et al., 2020;Moreira, 2021;Palmeira et al., 2021). Historically, the Psychiatric Reform movement posed the redefinition of the meaning of autonomy for community based mental health service users as a clinical-political challenge. This meaning of autonomy must broaden and even shift the meaning inaugurated by modernity, because autonomy is no longer conceived as strictly individual. In the Brazilian Psychiatric Reform movement, the process of becoming autonomous and of emancipation is considered collective and shared. (Gonçalves andCampos, 2017, p. 1545, our translation) Obstacles to full participation Participation in research, treatment, and societal life were intertwined in most studies and not analyzed separately. Key obstacles to full participation were conceptions of mental health (Jorge et al., 2012;Serpa Junior et al., 2014) (e.g., the biomedical model and psychopathology); systemic issues (Emerich et al., 2014;Palmeira et al., 2021) (e.g., tutelage, violence, poverty, and lack of access to healthcare and basic rights); power asymmetries (Onocko Campos et al., 2012;Gonçalves and Campos, 2017;Moreira and Onocko-Campos, 2017) (e.g., primacy of academic and professional knowledge, infantilizing service users, and disenfranchisement in treatment). Consequently, the team's actions often are directed to the need for symptomatic remission, and their service users' words remain muted due to the consideration given to their symptoms. This suggests that the responses indicated by the teams are still supported by the medical-biological perspective of understanding the phenomena of mental suffering, which does not seem to match what should be the object of work in this new context: the existencedistress in relation with the social. (Moreira and Onocko-Campos, 2017, p. 471) --- Discussion The overarching aim of this review was to chart and synthesize the participatory research in mental health in Brazil. We identified 20 relevant studies. Studies stressed the importance of participation in research as part of a broader democratizing process, reshaping power and knowledge relationships between expert and experiential knowledge. Studies noted that empowerment and autonomy are at the center of the Brazilian Psychiatric Reform movement and that participatory research lends itself to support this emancipatory project. Included studies highlighted that Brazilian mental health service users endure intersecting and synergistic processes of social exclusion that must be acknowledged. The biomedical model's reductionist views of mental health, violence, poverty, and social exclusion were all identified as barriers to full participation in research and in shaping public policy. Overall, Brazilian researchers did not define participatory research as distinct from other important participatory processes in society, including mental health service evaluation, public policy, advocacy, and broader claims of rights and liberties citizens must be entitled to. By refusing to treat these domains separately, Brazilian researchers have emphasized service evaluation and qualitative research in lieu of efficacy and effectiveness trials to establish the evidence base of specific interventions. This may be the result of the historical partnership of mental health professionals who consider themselves militants and advocates for the rights of service users. This configuration is not as common in the Global North, which tends to place service users and providers on opposite political sides with conflicting interests. Brazilian participatory research has been gradually developing in the country, and included studies highlighted the importance of participation; however, most studies only included participants in member checking activities. Claims of full participation were not substantiated or well described, and authorship was limited to researchers, even when there were claims of a participatory writing process. None of the studies reported full demographic information (age, sex/gender, and race/ethnicity), suggesting that reporting practices are inconsistent throughout and are not specific to participation. Participation is further complicated by multiple vulnerabilities mental health service users experience in Brazil, including poverty, limited literacy, and social exclusion. The tension between special and universal rights, a perhaps false tension if we consider the United Nations Convention on the Rights of Persons with Disabilities (CRPD) and a rights-based approach to mental health (P<unk>ras, 2017), featured in included studies. The emphasis of this literature on the concepts of autonomy and empowerment challenges the paternalistic nature of special benefits as they often rely on specific diagnoses and imply a deficit, perpetuating the narrow biomedical and individual-centered understanding of disabilities. On the other hand, special rights have historically played an important role in overcoming past and ongoing oppression and exclusion of this population promoting social justice. We note that nearly half of the included studies were conducted in partnership with a Global North country, suggesting that native participatory experiences not mediated by HICs are even rarer. The existence of well-developed theories of participatory research in Brazil using local epistemologies suggests that the scarcity of this type of research may be more related to funding inequities than a deficit in the field. Since 1988, participation in public policy development and broader political participation in Brazil has been expanding, especially in public health. In mental health, participatory service evaluation is one of the most developed areas in which participation has been more fully implemented (Ricci et al., 2020); nevertheless, projects still fail to center service users and often only include other stakeholders (Furtado et al., 2013) (e.g., policymakers, administrators, providers, and leadership). Compared to HICs, Brazil still lags in this area. This must be understood within the broader landscape of global mental health funding inequities and consider that 98% of all mental health research funding comes from HICs (Woelbert et al., 2020). A systematic review of the literature shows that mental health research in Brazil places a bigger focus on providers and work processes rather than outcomes and service users' perspectives (de Rosalmeida Dantas and Raimundo Oda, 2014). Although the Psychiatric Reform movement has promoted a shift in the traditional psychiatric paradigm, the biomedical model of psychiatry is still pervasive in community-based mental health services (Jorge et al., 2012;Serpa Junior et al., 2014); thus, power asymmetries in care continue, further supporting the need for increased participation. Influences from HICs appear in how participation is understood and practiced in mental health research in Brazil (notably in the use of fourth-generation evaluation methods); however, it is safe to say that the history of the Sanitary and Psychiatric reform movements in the country and the local scholarly traditions shaped a well-defined autochthonous field. Participation in Brazil does not fit neatly into conceptual definitions from the Global North, such as user-led (Rose, 2003) and co-creation (Greenhalgh et al., 2016); instead, Brazilian authors stress that participation has multiple meanings and applications. Importantly, participation is strongly rooted in what Brazilian authors call collective autonomy and in the rights of those who face multiple vulnerabilities and social exclusion (Passos et al., 2013). While tensions between research and activism are common (Montenegro, 2018), the neat separation of interests between providers and service users is not relevant in the Brazilian context, and partnerships across stakeholders are fundamental. It is noteworthy that 9 out of the 20 selected studies reported on a partnership with a Canadian university. Globally, mental health funding is unevenly distributed with HICs setting the agenda for the rest of the world. Along with setting the agenda, partnerships between the Global North and Global South in research are likely to be asymmetric despite good intentions. The project that yielded almost half of the publications we report here involved the translation and adaptation of a Canadian instrument, and not the creation of a native instrument. Such partnerships rarely ensure sustainability, and the impact of such projects tends to be limited to the life of the grant. Our study team has had multiple experiences with progressive researchers from the Global North who refuse to acknowledge a partnership between mental health providers and service users is possible. By setting the terms of what participation means, and what counts and does not, the Global North effectively continues to erase our history and deny our entrance in the debate unless we do so from a place of need and helplessness. --- Limitations Our study had limitations. Our group had to build creative strategies to overcome challenges related to missing metadata and unorthodox scientific reporting practices. The definitions of empirical research do not fully map out to how research is conducted in Brazil, making it difficult to extract data using traditional methods. Our team excluded theses and numerous articles because it was not clear if they were empirical studies or not, even though some presented data. Studies using cartographic methods were largely excluded, though many claimed to be participatory. While reviews of the scientific literature are important, they fail to include other forms of knowledge production and participation. This is particularly true for the Global South and relates to the inequities mentioned in our discussion. --- Conclusions Our study reveals important knowledge gaps. Participatory procedures were generally not well described except for the narrative data validation focus groups which are clearly conceptualized and form a discrete body of literature with well-described methods and Cambridge Prisms: Global Mental Health procedures. During the screening process, this became evident to our group, and we were forced to exclude potentially relevant articles because methods were not well described or were not described at all. Given the importance of participation in mental health research, the field could benefit from more standardized ways of reporting participatory procedures, which would in turn create better accountability of what counts as meaningful participation. We must acknowledge that the scientific literature does not fully capture the wealth of knowledge production and participation in the country. New and creative methods to synthesize the scientific and non-scientific (e.g., artistic) knowledge base produced by mental health service users are needed. Among the numerous obstacles to participation, disenfranchisement, poverty, and lack of access to social and civil rights are particularly relevant. The intersecting vulnerabilities of being Brazilian and diagnosed with mental illness substantively impact participation in research and beyond. Service users who rely on disability benefits to survive and advocate for universal civil rights exist in a paradox that cannot be sorted through research methodologies but forces us instead to contend with the political nature of research that is committed to social change. Reparations for coloniality in global mental health research are due. Should the Global North be truly interested in leveling historical asymmetries and inequalities in the world, the first step should be the equitable distribution of research funding. HIC researchers must enter partnerships with the Global South from a place of curiosity and solidarity. Capacity building should be grounded in mutuality instead of technology/knowledge transfer. Participatory research in Brazil is rich and original. The challenges for its expansion and full implementation must be understood within a broader social context of disenfranchisement, poverty, and lack of fundamental rights. This field holds true to the Latin American origins of participation as a transformative and democratic exercise and to the tenets of the Psychiatric Reform movement that questioned the primacy of the biomedical model in mental health, exposed its contradictions, and paved the way to a community-based network of services that centered dignity and freedom as inalienable rights of every citizen. We hope to see participatory research in Brazil expand and flourish as it has in HICs, the rhizomes already exist. Open peer review. To view the open peer review materials for this article, please visit http://doi.org/10.1017/gmh.2023.12. --- Data availability statement. Data availability is not applicable to this review article as no new data were created or analyzed in this study. --- Were participants co-authors? Alves et al. (2018) Fourth-generation evaluationimportance of stakeholder engagement. Validation of results in focus group sessions. No Chassot and da Silva (2018) Participatory intervention research: a confluence of influences including Brazilian health service stakeholder involvement tradition, health and mental health participatory research, and intervention research. Participants were co-researchers and participated in all stages. No Emerich et al. (2014) Fourth-generation evaluationimportance of stakeholder engagement. Narrative validation through hermeneutic focus groups. No Garcia et al. (2017) Problematizing the relationship between academic and popular/ community knowledge in the public health field. Leveling the dialogue across disciplines. Attending groups at the university Unclear Gonçalves and Campos (2017) Use of narratives as a means to access experience, not turning voices into objects, researching with and not about. Narrative validation through hermeneutic focus groups. No Jorge et al. (2012) Centering the lived experience of participants of the Autonomous Management Groups with an emphasis on the experiences between service users and their provider team; fourth-generation evaluation. Narrative validation through hermeneutic focus groups. No Kantorski et al. (2009) Qualitative evaluation participatory research, supported by the Gadamerian hermeneutics. Narrative validation through hermeneutic focus groups. No Lima et al. (2014) Partnership between researchers and providers and family members. Highlighting the importance of participation in the process of public policy formulation and the relationship between research and action. Workshop to jointly build the final measures. No Massa and Moreira (2019) To look for and respect the meanings that participants attribute to the studied phenomena and understanding research as a way to produce knowledge. Co-leading intervention groups during the implementation process. No Senna and Azambuja (2019) Balancing power differentials between researchers and participants. Research with and not about participants. Participants were lecturers at the university. No Serpa Junior et al. (2014) Using narratives to access the subjective experience and biography of participants. Service users learning about how psychiatrists think and vice versa. Participants validated each other's narratives through focus groups. No Silveira

Participatory research denotes the engagement and meaningful involvement of the community of interest across multiple stages of investigation, from design to data collection, analysis, and publication. Traditionally, people with first-hand experience of psychiatric diagnoses, service users, and those living with a psychosocial disability have been seen objects rather than agents of research and knowledge production, despite the ethical and practical benefits of their involvement. The state of the art of knowledge about participatory research in mental health Brazil is poorly understood outside of its local context. The purpose of this article was to conduct a scoping review of participatory and user-led research in mental health in Brazil. We identified 20 articles that met eligibility criteria. Participation in research was not treated as separate from participation in shaping mental health policy, driving care, or the broader right to fully participate in societal life and enjoy social and civil rights. Studies identified several obstacles to full participation, including the biomedical model, primacy of academic and scientific knowledge, and systemic barriers. Our extraction, charting, and synthesis yielded four themes: power, knowledge, autonomy, and empowerment. Participation in this context must address the intersecting vulnerabilities experienced by those who are both Brazilian and labeled as having a mental illness. Participatory research and Global South leadership must foreground local epistemologies that can contribute to the global debate about participation and mental health research.Our scoping review of participatory and user-led research in mental health in Brazil highlights the importance of engaging the community of interest in the research process and challenges the traditional view of people with psychiatric diagnoses as mere research objects. Our findings reveal that participatory research in mental health in Brazil is not treated as separate from participation in shaping mental health policy, driving care, or the broader right to fully participate in societal life and enjoy social and civil rights. We identified several obstacles to full participation, including the biomedical model, primacy of academic and scientific knowledge, and systemic barriers. By foregrounding local epistemologies and promoting Global South leadership in mental health research, our work contributes to the global debate about participation and mental health research. Our findings have implications for mental health research in Brazil and beyond, and we anticipate that our research will be used to inform the development of more inclusive and equitable research practices in mental health worldwide.

an emphasis on the experiences between service users and their provider team; fourth-generation evaluation. Narrative validation through hermeneutic focus groups. No Kantorski et al. (2009) Qualitative evaluation participatory research, supported by the Gadamerian hermeneutics. Narrative validation through hermeneutic focus groups. No Lima et al. (2014) Partnership between researchers and providers and family members. Highlighting the importance of participation in the process of public policy formulation and the relationship between research and action. Workshop to jointly build the final measures. No Massa and Moreira (2019) To look for and respect the meanings that participants attribute to the studied phenomena and understanding research as a way to produce knowledge. Co-leading intervention groups during the implementation process. No Senna and Azambuja (2019) Balancing power differentials between researchers and participants. Research with and not about participants. Participants were lecturers at the university. No Serpa Junior et al. (2014) Using narratives to access the subjective experience and biography of participants. Service users learning about how psychiatrists think and vice versa. Participants validated each other's narratives through focus groups. No Silveira et al. (2014) Balance the distribution of expertise so that knowledge is not exclusively with the researcher. In this perspective, research actively involves everyone in a transformation process. Service users wrote about the research experience. No Vaz et al. (2019) Strengthen service user protagonism and increase participation of non-traditional agents in jointly creating dialogic knowledge. Overcome the subject-object dichotomy and its assumed scientific objectivity. Stakeholders were invited to participate at the design, data collection, data analysis, and recommendation stages. Financial support. This study was supported by the Foundation for Excellence in Mental Health Care. --- No Competing interest. The authors declare no competing interests exist.

Participatory research denotes the engagement and meaningful involvement of the community of interest across multiple stages of investigation, from design to data collection, analysis, and publication. Traditionally, people with first-hand experience of psychiatric diagnoses, service users, and those living with a psychosocial disability have been seen objects rather than agents of research and knowledge production, despite the ethical and practical benefits of their involvement. The state of the art of knowledge about participatory research in mental health Brazil is poorly understood outside of its local context. The purpose of this article was to conduct a scoping review of participatory and user-led research in mental health in Brazil. We identified 20 articles that met eligibility criteria. Participation in research was not treated as separate from participation in shaping mental health policy, driving care, or the broader right to fully participate in societal life and enjoy social and civil rights. Studies identified several obstacles to full participation, including the biomedical model, primacy of academic and scientific knowledge, and systemic barriers. Our extraction, charting, and synthesis yielded four themes: power, knowledge, autonomy, and empowerment. Participation in this context must address the intersecting vulnerabilities experienced by those who are both Brazilian and labeled as having a mental illness. Participatory research and Global South leadership must foreground local epistemologies that can contribute to the global debate about participation and mental health research.Our scoping review of participatory and user-led research in mental health in Brazil highlights the importance of engaging the community of interest in the research process and challenges the traditional view of people with psychiatric diagnoses as mere research objects. Our findings reveal that participatory research in mental health in Brazil is not treated as separate from participation in shaping mental health policy, driving care, or the broader right to fully participate in societal life and enjoy social and civil rights. We identified several obstacles to full participation, including the biomedical model, primacy of academic and scientific knowledge, and systemic barriers. By foregrounding local epistemologies and promoting Global South leadership in mental health research, our work contributes to the global debate about participation and mental health research. Our findings have implications for mental health research in Brazil and beyond, and we anticipate that our research will be used to inform the development of more inclusive and equitable research practices in mental health worldwide.

Introduction Present socio-economical changes in our society require its members to play many roles and tasks. A good state of health and effective functioning are important not only for individual needs, but also for group needs. Middle-aged people constitute a considerable part of our modern society and middle-aged women have to perform family, professional and social duties, which leads to time pressure and overburden of role duties. It is recognized that the mechanism of taking the responsibility for the social role performed is stronger in women than in men. Karren Haas attributes more responsibility bound with home duties to women than to men [1]. This is why contemporary women have to combine professional tasks with a family life. Their knowledge and life experience make them undertake social activity eagerly. At the same time middle-aged women experience hormonal changes of menopausal transition. The menopausal transition as one of the stages in women's life can bring many physical and psychological ailments which complicate the accomplishment of family, professional and social tasks. Climacteric symptoms can lessen women's psychophysical skills and cause worse role functioning and eventually a decline in quality of life [2]. The average lifespan of a Polish woman is 80 years and the last menstrual period occurs on average in the 50 th year of age. So almost one third of a woman's life occurs after menopause. By any means it is not the twilight of the life, but a subsequent stage of existence, to which every woman should prepare herself earlier and pass through in good health [3]. The menopausal transition is a physiological stage in women's life and affects all women. Climacteric disorders can last for a few years [4] and can lead to disorders in family relations and occupational disturbances. The International Menopause Day is celebrated on 18 October. It recalls that this stage does not have to be a period of suffering and renunciations, but thanks to contemporary medicine achievements the menopausal aliments can be easily alleviated and some of Professional activity of Polish peri-and postmenopausal women them even eliminated completely. Especially on that day we should emphasise that regular check-ups, wellbalanced diet and active lifestyle are very important for maintaining women's good health and well-being [3]. Nowadays women occupy job positions which require taking responsibility and the ability of taking important decisions quickly. Unfortunately these tasks may be difficult to perform by peri-and postmenopausal women as these women may be suffering from memory or concentration disorders. For these reasons it is very important to help peri-and postmenopausal women in their professional functioning [5]. Professionally active climacteric women can have a difficult time performing their jobs. Typical climacteric signs as heavy sweating, hot flashes, vertigo, mood swings, irritability can obstruct fulfilling professional tasks. The stereotype of an old, snappish female clerk is well known to all of us. Stereotypes and ideas have their reasons. The menopause symptoms are some of them. It happens that women, who were feisty and selfconfident, start to doubt their skills during the menopausal transition. Simultaneously occupational burnout and the feeling of routine may appear. Women start to lose their professional stability and to be afraid of losing their jobs. They feel the breaths of younger colleagues on their backs. One cannot allow to be overcome by depressing mood, especially when working. Stress-free atmosphere in the office is very important. Offices and company premises are often not adjusted to mature women's needs. They lack air conditioning and adequate ventilation. There are no separated areas where older employees could have some rest and relax when a crisis situation occurs. This is why women are afraid to admit they suffer from menopausal ailments. Although they do not feel comfortable in their jobs they do not inform their employers about it. Menopausal women are ashamed to acknowledge experiencing these bothersome symptoms. It leads to a vicious cycle mechanism and worsens the situation. It may be worth starting an open discussion about menopausal women's problems at their working place and create optimal working conditions for them as it has been done for pregnant and breastfeeding women [6]. An attempt to evaluate peri-and postmenopausal women's professional status and job satisfaction is very important from a medical and psychological point of view [5]. Women often work in different professions than men, they differ from men in terms of salary and working hours. These differences result from education, preoccupation with family life or even life expectations. The level of economical activity is lower in women than in men. In 2013, in Poland we had 554 economically passive men per 1000 economically active women, whereas 1060 economically passive women for 1000 economically active men. We can observe feminization of economically inactive population -women constitute over 61% of that population. The distribution of professional activity according to age is different among men and women. The highest percentage of professionally active women is in the 40-44 age group, whereas among men in the 35-39 age group. There are clear differences in the structure of the population of working men and women according to their education. The majority of working women have higher education, then secondary education. On the contrary, men with secondary and vocational education dominate in the population of working men. The unemployment rate is higher among women than among men. It is difficult to find a job especially for women coming back to a labour market after a longer break or seeking for their first job. In Poland in every age group except the age group over 55 years, the unemployment rate is higher among women than among men. In every occupational group men earn more than women. In October 2012, the biggest differences in average gross salary between men and women were noted in the group of industrial workers and craftsmen. The average women's salary constituted 66.3% of the average men's salary. The smallest differences were spotted in the occupational group of office workers, where the average women's salary constituted 98.1% of the average men's salary. It is more difficult for women to compete on the labour market because some employers estimate women's work as less valuable. The lower level of seniority may be the consequence of this disparity [7]. The occupational situation of mature women is determined by many barriers and stereotypes, which impede women's access to work and professional promotion. These barriers involve many spheres of life: mental, institutional and individual. They refer both to women and other participants of the labour market: the state, employers and co-workers. An extremely important aspect of women's presence on the job market is the mentality and a lot of gender stereotypes which women have to face and which mainly determine the position of women on the labour market. The promotion of equality in politics, business and social life will have a positive influence on equal perception and equal treatment. More women in politics or business means better understanding of women's needs and also better quality of law created and solutions implemented, because the point of view of both sexes is taken into consideration. Women have to continue their education, struggle for better treatment at work only because the stereotype of "the weaker sex" still exists in our mentality. The aim of the study was to estimate the professional status and job satisfaction of Polish women aged 40-65. --- Material and methods The research was carried out by means of a survey method, a poll technique. The research was conducted between January and March 2016. 90 women aged 45-65 (arithmetic mean 51.93, standard deviation 5.47), inhabitants of the Lublin Voivodeship constituted the studied group. An original questionnaire consisting of 35 questions was used as a research tool. The questionnaire included closed questions, multiple choice questions and some open questions. All respondents were asked about their age, marital status, education and having children. The gathered material was subjected to descriptive and statistical analysis. Chi 2 test, U Mann-Whitney test, Kruskal-Wallis test and Spearman correlation coefficient were used. A statistical significance level was set as <unk> = 0.05. Almost a half of the respondents (48%) lived in cities with over 100,000 inhabitants, 30% of women lived in towns inhabited by fewer than 100,000 people. The remaining 22% of respondents were country dwellers. Married women constituted the majority of the studied group (57%). 17% of the polled women were widows and 12% were divorced. The percentages of unmarried women and respondents currently in an informal relationship were equal (both 6%). Women presently in separation constituted the smallest group (3% of the respondents). More than one third of the examined women (34%) had secondary education. In turn, women with a master's degree constituted 22% of the polled group, women with a bachelor's degree 18% of the examined and 21% of respondents had vocational education. 43% of the polled group were mothers of adult children, who had already left the family home, 41% of the women still lived with their children. The percentage of childless women was 16%. --- Results The differences between the respondents' age and their professional status are visible. The arithmetic mean of the age of women in full-time employment was 51.84 years whereas the arithmetic mean of the age of women in part-time or seasonal employment was 52.64. No statistically significant differences were found between full-time employees and part-time or seasonal employees according to the age. These data are displayed in Table I. There was no linear correlation between age and job satisfaction found (<unk> Spearman = 0.000, p = 0.997). It turned out that women aged 45-60 years showed statistically lower job satisfaction (M = 6.41) than women aged above 60 (M = 7.30). These data are shown in Table II. The permanent place of residence influenced professional status. The highest percentage of respondents in full-time employment lived in large cities. On the other hand, the smalest number of full-time employees lived in the country. In the studied group there was only one woman living in a small town and working part time or performing seasonal work. These data are presented in Table III. We have found no statistical differences between the respondents' permanent place of residence and their job satisfaction, what is depicted in Table IV It is interesting that only 10% of women in a very good financial situation were full-time employees and 32% of women who had estimated their financial situation as 'good' worked full time. 36% of examined women in an average financial situation worked full time and only 9% worked part time or performed seasonal work. These results show that professional status affects the women's financial situation because full-time employees estimated their financial situation as better than women working part time or from time to time. These relations are depicted in Table V. However, no statistical differences were found between the examined women's professional status and their financial situation (<unk> 2 = 2.354, p = 0.671). The financial situation of working women determined their job satisfaction. Well-off respondents were more pleased with their jobs than women in a worse financial situation. Rho Spearman correlation coefficient showed a highly statistically significant, moderately strong correlation between the respondents' financial situation and their job satisfaction. The better the financial situation the higher job satisfaction. Apart from that, it turned out that group 1 that is the respondents with a very good financial situation were characterised by a statistically higher level of job satisfaction (M = 8.50) than women from group 3 and 4 that is with an average (M = 6.00) and bad (M = 4.00) financial situation. These data are shown in Table VI. We found no statistical differences between the respondents in full-time employment and respondents in part-time or seasonal employment in terms of selfestimation of the health status (U Mann-Whitney Test: Z = -0.358, p = 0.721), what is presented in Table VII. The respondents' education affected their self-estimation of the health status. Women with vocational and secondary education evaluated their health status as 'good' (13% of both women with vocational and secondary education) and 'I experience some minor ailments from time to time' (7% and 16%, respectively). These results are presented in Table VIII. There is no statistically significant linear correlation (p = 0.320, Spearman's rho = 0.106) between self-assessment of the health status and the level of education. --- Discussion Menopause is a very difficult period of time for each woman. In these years women experience not only physical changes in their body, but also changes in their psyche, which can influence their professional and family life [8]. Many doctors emphasize that professional activity, fulfilling everyday duties and meeting friends may alleviate the symptoms of menopause. Some women do not even notice the symptoms of menopause [9]. The symptoms may be also reduced by a healthy lifestyle, a healthy diet, regular exercise, maintaining the appropriate body mass and prevention measures [10]. In our study we estimated the professional activity and work satisfaction of women aged 45-65. The Eurostat statistical data from 2014 concerning the employment in the European Union show there are differences in the number of working hours between genders. In 2014, almost one third of women aged 15-65 were working part time. The employment rates are lower for women and for the elderly [11]. Our data suggest that the arithmetic mean of the age of women in full-time employment was 51.84 years whereas the arithmetic mean of the age of women in part-time or seasonal employment was 52.64. The discrimination in the job market because of gender and age still exists in social mentality. The results obtained by Schulz and co-authors prove that job satisfaction increases with age. Older employees were more pleased with their jobs than younger ones, who just had started their professional activity [12]. Results of our studies indicate that women aged 45-60 presented significantly lower job satisfaction than respondents over 60 years of age. So we can draw a conclusion that the job satisfaction depends on the age of the employees. The 2013 CBOS research revealed that general work satisfaction was affected by non-financial factors, whereas the level of remuneration and the feeling of professional stability played a lesser role [13]. On the other hand, our results proved that the financial situation of working women determined their job satisfaction. Welloff respondents were more pleased with their jobs than women in a worse financial situation. According to Gojd<unk> and co-workers' study describing the quality of life in Polish perimenopausal female doctors, there is a relationship between the self-estimation of the health status and having regular meals or the frequency of physical exercise [14]. So we can see the relationship between the health behaviours and self-estimation of the health status. In our studies no significant differences were found between the respondents' education and professional activity and their self-estimation of the health status. It is worth pointing out that perimenopausal women's professional activity has a positive influence on their self-esteem. It was confirmed by Czarnecka-Iwa<unk>czuk et al. who reported that perimenopausal women active on the job market had higher self-esteem than women receiving the benefits for the disabled or old age pensioners [15]. The 2015 CBOS research proved the relationship between the place of residence and the level of satisfaction from professional career. The smaller town dwellers were less satisfied with their career than inhabitants of larger cities. Moreover, the studies suggest that respondents from the 55-64 age group manifest dissatisfaction with their financial situation more often [16]. Our findings proved no relationship between the respondents' place of residence and their job satisfaction. As we can see from the above-mentioned data, professional status and job satisfaction of peri-and postmenopausal women are affected by various factors. Both our results and the review of the literature prove that the topic is very important. Current demographic statistics inform us about the rising percentage of the elderly in the society. The number of peri-and postmenopausal women in the society also increases constantly. The existing discrimination of women on the Polish job market and the 'weaker sex' stereotype lessen the role of women. Further research on the situation of mature Polish women on the labour market is necessary. --- Conclusions 1. The differences between the respondents' age and professional status were visible. The arithmetic mean of the age of women in full-time employment was 51.84 years whereas the arithmetic mean of the age of women in part-time or seasonal employment was 52.64. 2. The respondents' place of residence affected their professional status. The highest percentage of women in full-time employment lived in large cities. The relation was close to statistical significance. 3. The financial situation of working women determined their job satisfaction. Well-off respondents were more pleased with their jobs than women in a worse financial situation. --- Disclosure Authors report no conflict of interest.

The menopausal transition as one of the stages in women's life can bring many ailments which complicate the accomplishment of professional duties. The situation of women on the Polish job market is determined by many barriers and stereotypes which impede women's access to jobs and professional promotions. Material and methods: The aim of the study was to estimate the professional status and job satisfaction of women aged 40-65. The research was carried out by a survey method, a poll technique in the period from January to March 2016. An original questionnaire consisting of 35 questions was used as a research tool. Results: The analysis of obtained results shows that the arithmetic mean of the age of women in full-time employment was 51.84 years whereas the arithmetic mean of the age of women in part-time or seasonal employment was 52.64. Women aged 45-60 were characterised by a significantly lower job satisfaction than women over 60 years of age. The highest percentage of women in full-time employment lived in large cities. The financial situation of working women determined their job satisfaction. Well-off respondents were more pleased with their jobs than women in a worse financial situation. Conclusions: There is a strong need of education about counteraction of gender and age discrimination on the Polish labour market. The struggle against discrimination requires support on each level of the state administration through the implementation of numerous regulations facilitating the final period and resumption of women's professional activity.

INTRODUCTION The impact of the COVID-19 pandemic among pregnant women and their newborns, has not been extensively examined to date. 1 Some reports show an increase in maternal mortality during the pandemic. 2 3 In addition, some studies on mortality in pregnant women with COVID-19 have shown a higher risk of death among those with lower income and poorer living conditions. 4 5 Does this result from infection, poorer access to health services or other factors? Previous research on gender-based violence, congenital syphilis, 6 cardiovascular disease, diabetes and tuberculosis suggests that marginalisation and poverty increase health risks related to COVID-19 pandemic. 7 8 In Colombia, a study of health disparities in the COVID-19 mortality found that the risk of dying from COVID-19 is higher for indigenous people, those with public subsidised health insurance, and those from the lowest income group. 9 Several studies have addressed pregnancy complications and clinical care guidelines for pregnant women. 10 11 To date, however, there had not been an analysis of excess maternal mortality and related social inequities during COVID-19 pandemic. For this reason, the objective of this analysis was to examine the impact of the pandemic on maternal mortality and its implications for social inequities in Colombia. --- METHODS A descriptive and retrospective analysis was carried out implementing two methodologies: a time series analysis and the measurement --- STRENGTHS AND LIMITATIONS OF THIS STUDY <unk> This study's main strength was to apply a combination of quantitative methods from official epidemiological surveillance databases to generate evidence about the impact that COVID-19 had on maternal mortality in Colombia. <unk> This study further generated evidence about social health inequities among pregnant women in Colombia, in the context of the COVID-19 pandemic. <unk> This is an ecological study, for which individual measurements are not carried out and therefore direct causality cannot be inferred. <unk> Analysis was not disaggregated to Colombian jurisdictions, thus likely under-representing the level of local variations in the results. Open access of both simple and complex inequalities. Official data source was the Colombian epidemiological surveillance system (Sivigila, by its Spanish acronym), including maternal deaths registers reported during period 2008-2020 to Sivigila. The response variable was the maternal mortality cases and ratios in Colombian women and the covariate used to measure inequality corresponded to the municipality of residence of the pregnant woman classified according to the Multidimensional Poverty Indices (MPIs) for the country. The Sivigila collects information on events of interest in public health, including maternal mortality, with coverage in every municipality. 12 Despite errors in capturing information at the local level, 13 the National Institute of Health has strengthened the quality, timeliness and comprehensiveness of the data. 14 During the pandemic, surveillance system has kept track of the maternal mortality. For time series analysis the Box-Jenkins approach 15 16 was implemented to identify changes in the number of maternal deaths in 2020 compared with expected deaths using national surveillance data for 2008-2019 as a baseline. Next, we created both simple and complex inequality measures for 2019 and 2020. 17 Simple inequities metrics included absolute, relative and population attributable fraction (PAF). The PAF is a metric that approximate the proportion of maternal mortality that would be reduced if all pregnant women lived in places with the best socioeconomic conditions. The complex measurements included Slope Inequality Index (SII) and Relative Inequality Index (RII). Both are weighted linear regressions that quantify the absolute and relative change, respectively, of the maternal mortality ratios (MMRs) comparing lowest and highest social positions in the population and use MPI from the National Institute of Statistics (Departamento Administrativo Nacional de Estad<unk>stica-DANE). 18 MPI consists of five dimensions of poverty: household education levels, children and youth conditions, work, health and public services access, evaluated by 15 questions. MPI provides a cut-off to classify a home as multidimensionally poor when it meets the criteria for at least 5 of the 15 items. The MPI provides a cut-off point of 20% of the live births in each quintile to classify a household as multidimensionally poor when it meets the criteria for at las 5 of the 15 elements. --- Patient and public involvement As this research is based on a secondary data review, there was no patient or public involvement in the conceptualisation or operationalisation of the research. --- Data and data adjustments Maternal mortality is defined as the death of a woman while pregnant or within 42 days after the termination of pregnancy. Late maternal deaths are those that occur after 42 days, but within 1 year of termination of pregnancy. 19 The total number of maternal deaths also include coincident cases (ie, road traffic or violence deaths while pregnant). For the time series analysis, 2008-2020 maternal mortality surveillance data was drawn from Sivigila. 12 For the inequities analysis data for 2019 and 2020 was used, but for these 2 years the Sivigila only reported information on early maternal deaths. For prior years, the surveillance system reported all maternal deaths: early, late and coincident. 20 21 A proportional adjustment projection estimate by epidemiological week and municipality was used for 2019 and 2020 to estimate total maternal deaths based on the relationship of early to total maternal deaths for prior years. 20 21 Time series analysis For time series estimates data were grouped by month. Verification of data dependence using L-Jung Box and Box-Pierce tests showed that this was successful as no outliers were found. Verification of seasonality was carried out using the Kruskall-Wallis and Friedman tests, to verify whether they corresponded to seasonal (Seasonal Autoregressive Integrated Moving Average) or non-seasonal (Autoregressive Integrated Moving Average (ARIMA)) models (details in online supplemental material). Forecasting of the model was based on data for 2008-2019. The best model for the 12 months of 2020 was chosen using Akaike criterion, white noise and normality tests in the residuals as the L-Jung Box and Jarque-Bera tests. A graphical comparison was done showing both the original data series and the adjusted model. A percentage change in the number of deaths was calculated for the period of the pandemic. The percentage change in reported cases by month was examined for the entire period analysed. The information was represented in double-axis graphs with the number of cases that occurred per month and the estimated cases with 95% CIs, defined as: Percentage change % = number of observed cases -number of estimated cases number of estimated cases <unk> 100 --- Social inequities analysis Annual MMRs were calculated by groups of municipalities, sorted by MPI, as a ratio of the total number of maternal deaths compared with the number of live births notified to DANE for each year. The municipalities were grouped by population quintiles of MPI. Each quintile is equivalent to an accumulated 20% of the population (according to data from the 2018 census), with quintile one (Q1) being the most affluent and the last quintile (Q5) being the poorest. Relative and absolute gaps were measured between the population of the quintiles of the MPI for 2019 and 2020. To calculate the PAF of excess mortality due to economic status, average MMR were compared with those in the highest quintile: PAF = MMRaverage-MMR Q1 MMRaverage <unk> 100 The calculation of complex inequities metrics was carried out by calculating the SII, by a weighted linear regression (details in online supplemental material). The health indicator (MMR) was estimated by quintile groups of municipalities according to MPI levels. We calculated relative and cumulative frequencies, and the ridit which corresponds to the average accumulated frequency. 17 22 23 After having the results of the SII, the RII was estimated according to the equation proposed by Kunst y Mackenbach. 24 For inequities monitoring, the magnitude of change of the MMR for each inequality indicator (absolute, relative, SII and RII) was calculated based on equations proposed by Bacallao. 25 The magnitude of change measure was calculated giving a bigger weight to the gap reduction than the variation in MMR 25 (details in online supplemental material). A positive result of this measure would be a reduction in the gap between municipalities of higher and lower socioeconomic status. A negative result would indicate a rising gap between these municipalities. The RECORD checklist criteria were met (see online supplemental material). --- RESULTS --- A total of 6342 deaths among pregnant women in Colombia during 2008-2020 as a result of complications associated with pregnancy, childbirth or the puerperium --- Open access were reported to Sivigila. In 2019, the prepandemic year, Colombia reported 511 maternal deaths with an MMR of 80 maternal deaths per 100 000 live births, according to the National Surveillance System. The time series model estimated 483 deaths expected for 2020 considering the behaviour of 2019, however, 544 early and late maternal deaths were recorded (MMR of 87 per 100 000 live births). This represents an excess death rate of 12.6% (95% CI -21.4% to 95.7%) of deaths. The time series ARIMA analysis showed that the six parameters were statistically significant, and their residuals were normally distributed (online supplemental table S1). Comparison of the original data series and modelled series is found in figure 1. Excess maternal deaths in Colombia for 2020 were highest in July and August; these were the only months when excess deaths were statistically significant (97.4%, 95% CI 35.1% to 250.0% and August 87.8%, 95% CI 30.5% to 220.8%) (figure 2). For the years under study, mortality rates were highest in municipalities with the lowest social quintile scores (figure 3). An increase in rates for 2020 was identified for the absolute, relative and PAF metrics (table 1). The rate of maternal deaths was almost three times higher in the lowest quintile municipalities compared with the highest quintile municipalities (table 1). Regarding complex inequities metrics, for 2020 there was also an increase. According to the SII for this year, in the worst poverty quintile there were 138 additional maternal deaths for every 100 000 live births. According to the RII, maternal mortality was 3.5 times higher in the worst condition quintile. In both years the R 2 value of the inequality regression models exceeded 80%, which indicates a strong correlation between MMR and socioeconomic vulnerability. The gap in maternal death rates was even higher in 2020 than in 2019 (table 1). --- DISCUSSION The COVID-19 pandemic was associated with a rise in maternal deaths in Colombia. This rise was greatest in municipalities with the lowest wealth quintile, and the gap in mortality rates between high and low quintile municipalities also rose. We found an excess of maternal mortality of 12.6% (95% CI -21.4% to 95.7%) in 2020, but with statistical significance only for the months of July and August (97.4%, 95% CI 35.1% to 250.0% and 87.8%, 95% CI 30.5% to 220.8%, respectively). Statistical significance of the results for July and August may reflect heightened transmission during the first peak of the pandemic. 19 26 Historical data does not suggest a higher MMR during those months. These results are consistent with those reported elsewhere, where not only maternal deaths, but also maternal and neonatal morbidity suffered, and rising inequality Open access gaps occurred during the pandemic. [27][28][29][30] This may reflect socioeconomic disadvantage, 31 poorer access to basic services and barriers to access to healthcare. 32 33 In Colombia rural and indigenous women, those with subsidised health insurance, those with low incomes and those with low educational achievement have historically had higher maternal mortality rates. 20 In Colombia, poorer access to health services due to pandemic-related closures and redirection of services likely reduced access to timely care for pregnant women, 34 leading to an increase in fatal outcomes. [35][36][37] Lockdowns also may have worsened the basic living conditions of the most vulnerable women. For example, transportation limitations due to mobility restrictions may have delayed the timely use of health services by pregnant women in remote areas. Cultural reasons, beliefs, and the same fear of contagion in health facilities may also have made pregnant women more hesitant to use services, as observed for other health concerns. 3 34 37 The importance of closures, transportation changes and popular fears regarding the use of existing services have important policy implications for future emergencies. Closures and redirection of services can be carried out in ways that favour vulnerable populations better. Publicity about transportation networks can reassure people that planning is done with them in mind. Education and health promotion can emphasise the importance of basic services, helping to balance the fear of non-use with existing fears regarding the use of services. When risks increase, people need ongoing orientation to manage competing needs better. [38][39][40][41] All this can be considered in response plans now to support health and reduce inequities in the future. This study has limitations. First, as an ecological analysis, direct causality cannot be inferred from population level data. Second, analysis was not disaggregated to individual Colombian departments or municipalities and thus local variations in results are not captured. Third, there was no testing or adjustment for the quality of Sivigila data. These data are considered to be of good overall quality and to have improved in recent years. 13 Sivigila complies with the standards of the Web-Based Maternal Mortality Epidemiological Surveillance System, 42 to assess under registration and data is carefully compared with National Administrative Department of Statistics (DANE) 43 software that verify the WHO recommendations to reduce the under-registration of maternal deaths. In Colombia, Sivigila information has also been widely used to calculate health inequities, among other analysis. 9 44 Finally, the results from this study in Colombia may not be generalised to those in countries with different cultural and economic systems and health service characteristics. --- CONCLUSION This study tracked maternal mortality trends during the first year of COVID-19 pandemic in Colombia. We identified both an increase in maternal deaths and a widening of inequity in deaths by socioeconomic level. The study highlights the need to monitor and proactively respond to known inequities that generate avoidable deaths. Maternal mortality is a key public health indicator for social development and equity in a society. Further monitoring during this and future disease outbreaks can be used to reduce inequities and improve health outcomes. The pandemic, appears associated with a delay in achieving the 2030 UN's Sustainable Development Goals (2030), related to health, well-being and reduction of maternal mortality. The 509 municipalities located in the poorest quintile in Colombia are not likely to reach the goal of reducing maternal mortality below 70 deaths per 100 000 live births. Although these municipalities are culturally and geographically heterogeneous, they share basic deficiencies in education, healthcare and access to basic services. We are challenged to overcome this reversal. 6 27 Contributors CC-O participated in conceptualisation, data analysis, writing and data validation, and is the guarantor of overall content. LHG participated in conceptualisation, data analysis, writing and editing. DD-J participated in methodology development, data validation, writing and editing. KC-C participated in writing, editing and data validation. RG participated in writing, editing and data validation. --- Competing interests None declared. Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research. --- Patient consent for publication Not applicable. Ethics approval Not applicable. Provenance and peer review Not commissioned; externally peer reviewed. Data availability statement Data are available upon reasonable request. Data used in the development of this manuscript may be shared on request. Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Objective The impact of the COVID-19 pandemic goes beyond morbidity and mortality from that disease. Increases in maternal mortality have also been described but have not been extensively studied to date. This study aimed to examine changes in maternal mortality and identify correlates and predictors of excess maternal mortality in Colombia during the pandemic. Setting Analysis of data from the national epidemiological surveillance databases of Colombia (Sivigila). Participants Deaths among 6342 Colombian pregnant women who experienced complications associated with pregnancy, childbirth or the perperium during 2008-2020 were included in this study. For inequalities analysis, a subsample of 1055 women from this group who died in 2019 or 2020 years were analysed. Methods We collected data from the national surveillance system (Sivigila) on maternal mortality. Analysis was carried out in two stages, starting with a time series modelling using the Box-Jenkins approach. Data from Sivigila for 2008-2019 were used to establish a baseline of expected mortality levels. Both simple and complex inequality metrics, with the maternal mortality ratios (MMRs), were then calculated using the Multidimensional Poverty Index as a socioeconomic proxy. Results Maternal deaths in 2020 were 12.6% (95% CI -21.4% to 95.7%) higher than expected. These excess deaths were statistically significant in elevation for the months of July (97.4%, 95% CI 35.1% to 250.0%) and August (87.8%, 95% CI 30.5% to 220.8%). The MMR was nearly three times higher in the poorest municipalities compared with the most affluent communities in 2020. Conclusions The COVID-19 pandemic had considerable impact on maternal health, not only by leading to increased deaths, but also by increasing social health inequity. Barriers to access and usage of essential health services are a challenge to achieving health-related Sustainable Development Goals.

Introduction The monkeypox virus (MPXV), a member of the POXVIRIDAE family of the genus Orthopoxvirus, is the pathogen responsible for monkeypox (mpox) fever, a zoonotic infectious disease [1]. The first known outbreak of the virus in monkeys was discovered in 1958 at an animal research centre in Copenhagen, Denmark, giving rise to the term "monkeypox" [2]. MPXV was not recognized as a human virus until 1970 when the virus was isolated from a child in the Democratic Republic of the Congo (DRC) who had symptoms of suspected smallpox infection [3]. The potential modes of MPXV transmission are either animal-human or human-human. Animal-human transmission is associated with direct contact with the blood, body fluids, and cutaneous or mucosal wounds of infected animal hosts [4]. Close contact with respiratory secretions, infected patient objects or surroundings, an infected person's skin lesions, and sexual contact with an infected person who has genital and groin lesions can cause human-tohuman transmission [5]. However, in the recent multi-country MPXV outbreak, the majority of MPXV cases have primarily been reported among men who have had sex with men (MSM) [6][7][8]. Human mpox has clinical features that are strikingly similar to those of smallpox, but less severe [9,10]. After an incubation period of 10 to 14 days, the majority of patients experience prodromic symptoms, including fever, malaise, and enlarged lymph nodes. About 90% of patients infected with MPXV develop lymphadenopathy [11,12]. After an initial prodromal period, a centrifugally distributed maculopapular rash developed. Lesions develop from macules to papules, vesicles, pustules, and eventually crust over in two to four weeks [13]. Treatment of MPXV is mainly symptomatic. The smallpox vaccination is 85% effective in protecting against mpox [14]. However, Modified Vaccinia Ankara-Bavarian Nordic (MVA-BN), an authorized vaccine, and the antiviral medications tecovirimat and brincidofovir are not easily available [15,16]. In 2003, the United States reported the first MPXV outbreak in humans outside of Africa, with 70 cases, which was associated with the importation of exotic pets from Ghana [17]. Since then, the incidence in endemic areas has substantially increased, and sporadic, travel-related incidents are on exponential rise in non-endemic nations [13,18]. In September 2017, Nigeria reported a major outbreak of human MPXV, with about 228 suspected and 60 confirmed cases caused by the West African clade [19]. In the ongoing outbreak, the UK Health Security Agency reported the first mpox case on May 07, 2022. Since that time, the World Health Organization (WHO) has consistently recorded reports of mpox cases [20,21]. Mpox was declared a public health emergency of international concern by the WHO on July 23, 2022, due to an upsurge in cases [22]. From January 1 st through November 17 th, 2022, 80221 confirmed cases of MPXV via PCR and 52 mortalities have been reported to the WHO from 110 countries [23]. Although MPXV has been reported in other southeast Asian countries (India, Thailand, and Indonesia) [24,25], so far, Pakistan has not reported any cases of MPXV. Although, there were reports on two rarely occurring cases of the zoonotic mpox infection [26], but the reporting of mpox cases in Pakistan was false news, which was refuted by the National Institute of Health on May 24, 2022 [27]. According to the WHO, a lack of awareness about the virus is one of the obstacles in preventing the re-emergence of MPXV [28]. Adherence to prevention and control initiatives is important in restricting the dissemination of the virus, which is largely influenced by the community' cumulative knowledge, attitudes, and practices (KAP) [29,30]. A comprehensive strategic framework must be established and implemented to prevent the emergence and spread of MPXV in Pakistan. Our findings emphasize the need to investigate the KAP towards MPXV among the general population of Pakistan. The study's conclusions could assist public health officials in strengthening policies, increasing public awareness of the MPXV outbreak, and organizing future public health initiatives. --- Methods --- Study design, setting and data collection An analytical cross-sectional study was applied to assemble data from Pakistan's general population using a non-probability convenience sampling technique. Using shared Google forms links across WhatsApp, Instagram, Facebook, and Twitter, Pakistan's most widely used and accessible social media platforms, the research team developed a questionnaire with informed consent. Between July 5 th, 2022, to August 1 st, 2022, data were collected across all of Pakistan's provinces (i.e., Sindh, Punjab, Baluchistan, and Khyber Pakhtunkhwa). Individuals below the age of 18, non-Pakistanis, and those who were unable to understand the questionnaire language were excluded from the study. --- Sample size Pakistan's population, according to the most recent census, is 207.7 million [31]. Using a Raosoft sample size calculator online, minimum sample size of 752 was calculated using a 90% confidence interval, 50% of the population, a 3% margin of error, and a total population of 207.774 million. We ultimately obtained 1040 responses to include in the survey, which will help us get more accurate and reliable results. --- Questionnaire development The self-reported questionnaire was developed using an extensive literature review [16,32,33] and information was sorted out from the WHO and Centers for Disease Control and Prevention's official websites (CDC) [34,35]. This survey form was reviewed by a team comprising senior community medicine specialists, public health experts, and infectious disease researchers to improve some details for the pseudo-validity, relevance, completeness, and clarity of individual sections. After a detailed discussion and audit, the authors finalized the survey. Then, 80 people participated in pilot research to examine the reliability of the system. Cronbach's alpha was reported to be 0.76, and the results of the pilot study were not included in the final analysis. The beginning of questionnaire was introduced with a brief overview of the nature and purpose of the study, informed consent, and then segregated into four divisions to assess the socio-demographics of the participants, knowledge, attitudes, and practices. Six questions regarding age, gender, marital status, residence, education, and monthly income were included in the demographic section. The second section evaluated the knowledge of participants of mpox by asking 10 questions. Each score was graded as "1" (correct) and "0" (wrong) with the scores ranging from 0 to 10. The study participants who scored 7 or more were classified as having good knowledge. The third section assessed attitudes. Seven questions were tested by a 5-point Likert scale as follows: 1 ("strongly disagree"), 2 ("disagree"), 3 ("neutral"), 4 ("agree"), and 5 ("strongly agree"). The overall score for the attitude section ranged from 7 to 35, with an overall higher score depicting a more positive attitude. A score higher than or equal to 26 was described as a positive attitude. The last section assessed the practices. Six questions were tested with a 5-point Likert scale as follows: 1 ("strongly disagree"), 2 ("disagree"), 3 ("neutral"), 4 ("agree"), and 5 ("strongly agree"). The practice section score ranged from 6 to 30, with an overall higher score depicting a more positive practice. A score more than or equal to 22 was considered a positive practice. --- Ethical considerations The anonymity and confidentiality of participants were ensured during the survey. The research did not include names or other personal information. Before beginning the questionnaire, informed consent was taken from the sampled study participants. During the process of survey, participants can deny and withdraw at any time before submission. The ethical review committee of the Pir Abdul Qadir Shah Jeelani Institute of Medical Sciences in Gambat, Pakistan, approved the study protocol (Ref. No: IRB/22/13). --- Statistical analysis The statistical package for social sciences (SPSS) version 25 was used for data analysis (IBM). While categorical variables were expressed as frequencies and percentages, numerical variables were quantified as means and standard deviations. Based on the variables and the data type, inferential statistics were used. Univariate logistic regression analysis carried out a comparison of knowledge, attitude, and practices with socio-demographic variables. The statistical results were demonstrated as odds ratio (OR) with a 95% confidence interval (CI) and p-value. In multivariable analysis, all with p-value <unk>0.25 were included, and an adjusted odds ratio (aOR) was reported with independent association. A p-value of <unk>0.05 was considered statistically significant. --- Results --- Socio-demographic characteristics A total of 1040 participants were involved in the final analysis. Of the total, 61.4% (n = 639) were male, 61.8% (n = 643) fell in the age group 21-30 years, 79.5% (n = 827) were single. 79.5% (n = 789) of the respondents were residents of urban areas, and 57.2% (n = 595) achieved graduation level education (Table 1). --- General knowledge, attitudes, and practices about Mpox This study determined that 34.4% (n = 358) of the participants had good knowledge, while 65.6% (n = 682) had poor knowledge about MPXV. 47.7% (n = 496) respondents correctly answered that MPXV was first isolated in Congo, and 87% (n = 905) correctly identified the causative agent of mpox is virus. 65% (n = 683) knew about preventions and 64% (n = 668) were familiar with the symptoms. Only 30% (n = 312) of participants were familiar with the fact that the Smallpox vaccine is also effective against MPXV (Table 2). Regarding attitudes, 41.7% (n = 434) of the participants had positive attitudes. 48.6% (n = 505) agreed that MPXV is a fatal disease, and 38.3% (n = 398) agreed that MPXV can cause a pandemic like the Coronavirus disease 2019 (COVID-19). 44.6% (n = 464) believed that traveling from endemic areas (such as Africa, Nigeria, etc.) to non-endemic areas should be banned (Table 3). Assessment of practices showed that the majority of participants, 57.7% (n = 600), had good practices. About half (48.0%) of respondents agreed that infected patients should be isolated from others. 38.9% (n = 405) strongly agreed that there is a need for more awareness of MPXV in the general public. When asked whether the government of Pakistan is taking adequate steps to tackle MPXV, 27.0% (n = 281) were neutral to making the decision (Table 4). --- Factors of good knowledge, attitudes, and practices The regression analysis revealed that men had a higher odds ratio than women [Adjusted odds ratio (aOR) = 1.324; 95% Confidence interval (CI) = 1.001-1.752, p = 0.049]. Participants with an education level from primary up to matriculation (Class-10) were 0.22 times less likely to have good knowledge than those with intermediate education (Class-12) (aOR = 0.220; 95% CI = 0.074-0.653, p = 0.006) (Table 5). With respect to the different variables affecting attitudes regarding MPXV, and based on statistical values obtained by multivariate logistic regression, participants with an income of 50,000-100,000 Pakistani Rupees (PKR) <unk>225-450 USD<unk> had lower odds of positive attitudes compared to participants who had an income of more than 100,000 PKR <unk>450 USD<unk>, (aOR = 0.593; 95% CI = 0.384-0.917, p = 0.019), while those aged 31-40 years had a more positive attitude (aOR = 2.454; 95% CI = 1.099-5.483, p = 0.029) (Table 6). With regards to good practices, males had lower odds of good practices than females (aOR = 0.741; 95% CI = 0.572-0.959, p = 0.023). --- Discussion To the best of our knowledge, this is the first study carried out in Pakistan to assess the KAP towards MPXV among the general population. Based on our results, most participants had poor knowledge and attitudes but positive practices toward MPXV. Gender and education had a significant impact on knowledge and practices. While monthly income status had a significant impact on attitudes. Our study findings indicate that most study participants (65.6%) had poor knowledge about MPXV. This inference is consistent with other recent research that revealed that people generally in Saudi Arabia, Bangladesh, and the Kurdistan region of Iraq had little knowledge of MPXV [16,36,37]. As MPXV is endemic in tropical rainforest regions, Pakistan lies in a temperate climate zone, and there has been no previous exposure to the disease in the country, rendering the disease unknown among the general population [12,38]. The majority were unaware of the origin of MPXV in the DRC. People visiting the endemic regions should be aware of the locally endemic diseases, which will help in the adoption of different preventive and treatment strategies, such as MPXV and COVID-19, which are endemic in the DRC and China, respectively [39,40]. Thus, it is necessary to initiate awareness campaigns in Pakistan, Data presented as frequency (percentage), and average scores (mean) per total participants. SD = strongly disagree, D = Disagree, N = Neutral, A = Agree, SA = strongly agree. Practice section was assessed by giving a score of 1 to strongly disagree and 5 to strongly agree. A score of greater than or equal to 22 was regarded as good practice. https://doi.org/10.1371/journal.pone.0279952.t004 especially for professional travelers, the business community, and tourists to the endemic region, to prevent the spread of disease to non-endemic areas. However, participants in this study had good knowledge about the causative factor of mpox fever (87.0%), prevention of the disease (65.0%), and common symptoms (64%) (Table 7). Except for a few variations, smallpox, and mpox viruses belong to the same family, and many similarities are found in terms of disease presentation, spread, prevention, control, and management plans [41][42][43]. This might be one reason for being known common symptomatology of MPXV infection and preventive measures. But only 30% knew that the smallpox vaccine is also effective against MPXV. Vaccine against smallpox is considered beneficial for the prevention and control of MPXV [44,45]. Some analyses also depicted that the MPXV outbreak was due to a break in vaccination against smallpox [46,47]. Our study shows that most participants (58.3%) had a negative attitude toward MPXV, which is consistent with a study conducted on the general public's attitude toward mpox in the United States [48]. Good knowledge is not beneficial without applying that knowledge in practical life. Individual knowledge can be practically utilized to adopt a positive attitude by engaging people in awareness activities. Studies in Iran and Indonesia about avian influenza postulated similar aspects of the correlation of knowledge with attitudes [49,50]. Most respondents believed that MPXV is a fatal disease that can cause a pandemic like COVID-19. The COVID-19 pandemic hit many people hrad by evolving a lot of complications for them, and based on this, it might be possible they are considering MPXV to be fatal and a pandemic in the future [51,52]. Some studies depict the role of the fatality of disease over one's psychology as being the case with the Ebola virus disease [53,54]. However, little information is available regarding the connection between MPXV and psychological issues [55]. In our study, most respondents had good practices compared to knowledge and attitudes. The findings are consistent with the studies conducted on general physicians showing good practices [32,56]. In our study, most participants believed that infected patients should be isolated from others and that preventive measures such as face masks, hand washing, and quarantine would prevent the spread of MPXV. This might be due to the recent pandemic of COVID-19, during which taking precautionary measures was the sole mode of protection all over the world before the vaccine's introduction [57]. The other reason for good practices could be recent alerts from WHO and CDC. In agreement with the previous study [36], our study reported significantly good knowledge of MPXV among male participants. This may be connected to how MPXV is transmitted because the majority of cases in the most recent outbreak have mostly been reported among men who have had sex with men (MSM) [58,59]. According to a recent pooled meta-analysis, sexual contact is involved in more than 91% of cases [6]. A study reported that all (100%) of the 54 MPX cases that presented at one health center in the UK were MSM [60]. Study findings affirm that respondents with higher education levels had better knowledge than those participants with qualifications of primary to matriculation (Class-10). Other studies with similar findings also revealed that a greater level of education was related to having good knowledge [61,62]. Education has a key role in promoting healthy behaviors and gaining control over epidemics, by improving knowledge levels about any health-related case. Probably, this is because of more exposure to media, internet, and newspapers that provide easy access to knowledge and a higher maturity level. However, the association of age, residence, and monthly income with knowledge did not withstand the regression analysis in the current study. Nevertheless, our results showed that the population with a low income had a more positive attitude. This is not consistent with many studies that revealed having a higher socioeconomic position corresponds with having a more positive attitude [63,64]. People with higher socioeconomic status have additional opportunities for higher education, residing in urban areas, and exposure to mass media. The study findings demonstrate that females were shown to have more good practices compared to males. Similar findings were observed in other studies showing that females take more preventive measures than males [65]. Women claimed to be more responsible for their health, potentially related to risk perception and the gender bias that women are socialized to be more concerned about health issues than men. Mpox is an emerging virulent disease poses a serious threat to worldwide public health [42]. Given the substantial risks posed by MPXV and the lack of mpox vaccine, prevention strategies play a crucial role in minimizing infection rates and halting the disease's spread. It has been acknowledged that increasing awareness through campaigns and seminars, as well as embracing preventive strategies can halt infectious diseases like COVID-19, influenza, and dengue from becoming epidemics or pandemics [66][67][68]. Studies conducted to assess knowledge and attitudes about disease prevention and improve public health interventions are futile without public interest and collaboration [32,69]. This highlights how crucial it is for the general public to follow preventative and control measures based on their knowledge, attitudes, and practice (KAP). --- Study strengths and limitations While interpreting the study's findings, some limitations should be considered. First, there was a possibility of selection bias because the convenience sampling method was used. Second, the questionnaire was distributed through an online system using different social media platforms. Consequently, there is a possibility of bias as underprivileged populations would not have been able to participate, and results cannot be extended to the entire community. Third, this study is based on self-reported data, which might lead to internet surfing to acquire information for a difficult and technical question in the survey, resulting in a biased response to some questions. According to what they believe to be expected, participants may have responded positively to attitude and practice questions. Fourth, a recall bias during the survey cannot be ignored. Finally, results stratified by marital status and residence should be interpreted with caution, since the majority of the participants were unmarried and urban. However, the current study was strengthened by the large sample size. To the researchers' knowledge, this is the first study analyzing KAP toward MPXV, in the Pakistan's public. Therefore, the results of this study may help health authorities to implement effective policies to combat the emergence of MPXV in Pakistan. --- Conclusion In summary, the present study demonstrated inadequate knowledge and attitudes regarding MPXV among the Pakistani population. Our results showed that good knowledge could result in good attitudes and practices, which are crucial to minimizing the expanding impacts of the disease. The current outbreak of MPXV in non-endemic countries warrants strict epidemiological surveillance to restrict the further dissemination of the outbreak in other non-endemic countries. International organizations, national health authorities, national organizations, health care workers, and the media sector must be actively involved in the implementation of appropriate interventions for the awareness and prevention of MPXV. --- Supporting information --- S1 File. (SAV) --- Author Contributions Conceptualization: Hashaam Jamil.

Monkeypox (mpox), re-emerging zoonotic infectious disease, is striking the world with serious public health concerns, especially in non-endemic countries. The public's knowledge and attitude towards the monkeypox virus (MPXV) influence their adherence to preventive strategies. Therefore, we aimed to assess the public's knowledge, attitudes, and perceptions (KAP) of MPXV in Pakistan.We collected data for this cross-sectional study from 1040 participants via online selfreported questionnaire from July 5 th , 2022, to August 1 st , 2022. The questionnaire consisted of a total of 29 items in four sections, assessing socio-demographics, knowledge, attitudes, and practices regarding MPXV. The data were analyzed using IBM SPSS V.25, and factors associated with MPXV knowledge, attitude, and practices were identified by using logistic regression analyses.A total of 1040 participants were included. 61.4% were male, and 57.2% had graduation level education. Only 34.4% had good knowledge about MPXV, and 30% knew the effectiveness of the smallpox vaccine against MPXV. 41.7% had a positive attitude, 48.6% agreed that it is a fatal disease, and 44.6% were in favour of banning travel from endemic to

INTRODUCTION India has a population of 1.3 billion, about 2/3rd of which resides in rural areas (1), about 1 in 10 belong to one of the many scheduled tribes (2). Rural and tribal communities face a higher disease burden than others, but have much less access to healthcare (2). Poor access to healthcare for these communities is often on account of uneven distribution of qualified health providers between rural and urban areas; and between tribal and non-tribal areas. According to The Health Workforce in India report (2016), there are almost four times as many doctors and nurses in urban areas than rural areas (3). Moreover, high levels of absenteeism (4), low motivation marked by social and family isolation, poor educational opportunities for children, insufficient pay (5) affect quality of healthcare in these areas. Task-shifting is one of the strategies used to enhance availability of skilled human resources in rural areas. The World Health Organization (WHO) defines task-shifting as: "Specific tasks are moved, where appropriate, from highly qualified health workers to health workers with shorter training and fewer qualifications in order to make more efficient use of the available human resources for health" (6). An emerging task-shifting strategy is up-skilling of nurses to undertake larger responsibilities than their historically defined role of assisting a physician. Task-redistribution is used synonymously to task-shifting, and we have preferentially used the former nomenclature. Previous reviews highlight the effectiveness of up-skilling nurses in meeting healthcare demands amidst growing burden of diseases (7)(8)(9)(10). The reviews clearly establish that when skilled, mandated, and supported, nurses are as effective as doctors in delivering primary health care. Higher patient satisfaction outcomes (7,8) have been observed with nurses as first point of contact (11) and thus patients more likely to keep their followup appointments with the nurses (12). This satisfaction is also associated with greater engagement with the patient (13) through counseling, two-way communication and drawing in contextspecific real life connect (14). Nurse-led clinics (15) have been effective in providing specialized curative care and in improving patient outcomes. Rural Pathways refer to a multi-pronged approach taken to sustain efforts to improve health outcomes in rural areas. Rural pathways may be designed to address specific needs identified in the community like the scarcity of health workforce in rural areas. O'Sullivan et al. (16) drew evidence and came up with a comprehensive checklist for implementing rural pathways for rural workforce in low-and middle-income countries (LMICs). The checklist highlights several components that initiate from the community needs to health worker selection, training and up-skilling, accreditation, and recognition as well as monitoring of outcomes. Examples from LMICs which were key to the lessons drawn for the rural pathway checklist focussed largely on availability of physicians or on community volunteers' up-skilling to perform certain primary care functions in rural areas. A significant proportion of task-redistribution and nurse-led healthcare interventions have been implemented and tested in high-income countries. Also, most research focuses on assessing effectiveness outcomes of nurse-led interventions. However, there is limited understanding of what it takes to ensure availability, retention and ensuring high performance of nurses in delivering primary health care, especially in rural and tribal areas. Several systemic barriers such as hierarchies in the workforce and structural barriers and gender inequality pose as barriers in enabling nurses (most of whom are women), to assume leadership roles in countries like India (17,18). There is a critical need for exploring rural pathways that ensure availability, retention, and effectiveness of nurses as primary healthcare providers in rural areas in LMICs. Basic Healthcare Services (BHS), a not-for-profit organization that authors are associated with, runs a network of six primary healthcare clinics (called AMRIT Clinics) in rural and tribal areas of South Rajasthan, India. The clinics are managed by Primary healthcare (PHC) nurses, supported by a visiting and on-call primary care physician at one end, and community health workers and volunteers on the other. We analyzed our evidence and experience of running these Clinics to identify what steps along the rural pathways are required to ensure availability and effectiveness of PHC nurses in delivering primary healthcare in such areas. --- METHODOLOGY --- Context Southern region of Rajasthan is predominantly rural, most of which is inhabited by a population belonging to one or other scheduled tribes (19). The region has the lowest Human Development Index (HDI) of 0.5 (20). Only about 35% of women in these communities are literate (1). From as many as 45% households, at least one man migrates to cities in search of livelihood (21). Half of all children are underweight and a similar proportion of mothers are malnourished too (20). Absence of functional health facilities in the area due to high rates of absenteeism (4), poor availability of food (19), and distress migration have a detrimental impact on the health status of the population in these areas, especially women and children. Careseeking (22) among these households remain erratic with only 6% --- Pathway Framework To evaluate the effectiveness of strengthening capacities of PHC nurses, we adapted the existing rural pathway framework (16) and contextualized it (Figure 1). The components of the Pathway are described in Table 1. --- Methods for the Study To assess the effectiveness of the pathway and to evaluate the development outcomes, secondary data from three sources was --- Sources of data for the pathway components --- Sourcing Organization's Human Resource Information System Nurses' qualification, background and place of origin data. --- Training The number of trainings conducted and the subject of trainings --- Data on retention and attrition The duration spent by each nurse at the clinic and the reasons for leaving. --- Data on nurse's satisfaction Qualitative data was accessed from nurses' interviews conducted by a third party organization named "Start Up!" for need assessment of leadership training (interviews conducted in February 2020) Quotes suggestive of the themes identified in the interviews. --- Data on patients managed Organization's Health Management Information System (HMIS) (data retrieved for the years 2015-2019) The data on the different types of cases managed by the nurses and the doctors. analyzed. The sources of data for each pathway component are explained in Table 2. The quantitative data was analyzed in MS Excel, using pivot tables to segregate provider level data and health condition wise data. For qualitative data, interview verbatim from a previous source was accessed, translated, and analyzed while keeping the names of the respondents anonymous. Emerging themes were identified, and supporting quotes were elicited to substantiate the themes. --- RESULTS We analyzed information from above sources to describe and assess the different components of rural pathways for PHC nurses. --- Sourcing Nurses registered with Indian Nursing Council; with either a Bachelor of Science in Nursing (BSc Nursing) or a Diploma in General or Auxiliary Nursing and Midwifery (GNM or /ANM) were eligible for recruitment as PHC nurses in AMRIT Clinics. An affirmative action was taken to source them from rural and tribal areas, as they are more likely to live and stay in such areas and more empathetic to needs of the communities. A total of 40 PHC nurses have been recruited in AMRIT Clinics from September 2012 till May 2020. Of these, 93% belonged to rural areas and 83% of them come from similar tribal communities that they serve (Table 3). Eighty-two percent of them had a diploma in GNM and 12.5% had a diploma in ANM. Two (5%) were interns pursuing a GNM degree. --- Training All nurses recruited as PHC Nurses undergo a focused onemonth induction training, followed by 3-6 months intensive onthe-job training at clinics. This is followed by continuous nursing education, 2-3 days per month. The training themes span across clinical knowledge and skills, communication and other soft skills, and management skills. Each component also emphasizes on context-specific understanding of social and cultural determinants of health. In addition, weekly primary care physician's visits and bi-monthly visits by Nurse Mentors are also opportunities to train on new and refresh on older learnings. "There is a good balance between training on clinical issues and (exposure to) social realities we see" shares a Nurse Coordinator about the training imparted to them. --- Task Redistribution Task redistribution is a key component of the pathway that enables PHC Nurses to perform their duties effectively. In AMRIT Clinics, PHC Nurses are entrusted and credentialed to perform the primary curative functions, and enabled to do so through continuous training, standardized protocols, pointof-care diagnostics, and easy access to tele-consultation. PHC nurses reside in the Clinic villages and provide care on all days. Primary care physician lives in the nearby town, visits the clinics weekly on designated days, and is available for tele-consultation at all times. Redistributing the tasks of primary curative care between PHC nurses and the physician does not equate to shifting of responsibility: the responsibility is jointly shared by the two. "When a patient comes with a condition we are not certain, we call the consult the doctor on telephone. We also ask the patient to come on the designated day for doctor's visit after assessment. This is how wee ensure that no patient returns without receiving any counsel" says a Nurse about the patient management procedure. --- Support and Mentoring The organization (BHS) supports PHC Nurses to play their role and mentors them for their personal and professional growth. Physicians provide mentoring and tele-consultation support to the PHC Nurses. Regular monitoring of data and feedback is provided to the clinic teams for quality of care. Career advancement opportunities are provided to the PHC nurses: six of them have so far advanced to Nurse Coordinator position, and one of them has progressed to the Nurse Mentor position. The Nurse Mentor and a Clinic Associate supervise and mentor the PHC nurses individually through periodic onsite visits and through on-line support. The mentors identify opportunities for their personal and professional growth and assist them to tap those opportunities. Additionally, an accommodation with basic amenities such as water, electricity, and sanitation; a competitive salary (comparable to salary in city hospitals) and a well-equipped clinic with an amicable and respectful environment is made available to all of them. --- Pathway Outcomes --- Work Satisfaction and Motivation Although no formal investigation was conducted to assess the job satisfaction, an independent organization conducted qualitative interviews with six PHC nurses to assess their leadership potential and needs for training. Interviews revealed a high sense of pride and satisfaction they perceived in managing the patients independently, equivalent to that of a physician: "We used to work under the doctor in the clinic but here we ourselves, it feels nice that we are doing doctor's work also, " shares a PHC nurse at one of the clinics. Nurses appear to have a high level of patient-centeredness, as illustrated by the following quote by a nurse who was interviewed: "We do not work for money here; days merge in nights, but we try to ensure that the patients get better." High motivation appears to be also on account of the team spirit: "Our team of doctors and nurses have always been supportive which keeps me motivated." "We feel gratified when we see the change in the community, it makes us forget all the fatigue from the hardwork, " shares a nurse about her work in AMRIT Clinics. --- Retention A total of 38 PHC nurses have been recruited in AMRIT over the last 7 years. Of those, 20 left the job after spending a median of 15 months (IQR: 8-35 months) A further analysis on segregating the PHC nurses based on their place of origin and caste found that the three from urban background areas stayed for a median duration of just over 6 months, while those from rural areas stayed for a median duration of 16 months. Out of 38 PHC nurses employed so far, 20 left for various reasons. The two major reasons for leaving the job were: shifting to a government job (10), and for family reasons (10). The appeal of a "government" job is often what drives the PHC nurses' aspirations. As one of them shared: "If we get a government job then we can get a posting near our home. We are working with this thought that in government job the salary will be better and would help in supporting our family." While the salary at AMRIT Clinics is comparable to the best of private hospitals in cities, salaries and perception of job-security for government jobs is much higher. Table 4 shows the comparison between the remuneration and perks offered by the government, private hospitals, and AMRIT Clinics. The average age of PHC nurses who have worked at the clinics was found to be 27 years, with the age range being 21-38 years. Of the 38 PHC nurses, 30 are married and more than 20 of them were mothers often with more than one child. The "family" reasons for leaving often pertained to child bearing or child rearing roles and presence at home for carrying out other household chores, such as care of ailing parents-in-law. As one of the PHC nurses puts it: "We are staying away from our family. We cannot bring them here. I have not stayed with the family since my marriage." The reason for their families not staying with them is lack of employment opportunities for their spouses and a poor quality of schools in the villages. --- Performance A retrospective analysis of the records from 2015 to 2019 of 3 AMRIT Clinics saw 87,227 patient visits, of whom 55,825 (64%) were women and 14,829 (17%) were children. While PHC nurses managed most of reproductive and child health conditions and communicable diseases, visiting physicians managed most of non-communicable diseases (Table 5). A major indicator of quality of the service provided at the clinics is trust amongst the community. Sixty-three percent of the total patients visits made were repeat visits at reflect their perception of the care they receive at the clinic. The PHC nurses are also equipped with a protocol to manage patients with an array of conditions and also Standard Operating Procedures (SOPs) to ensure that the operation of the clinics are uniform and of high quality across all clinics. --- DISCUSSION Our analysis demonstrates that a rural workforce pathway approach can be effective in planning for and assessing the factors that contribute to availability, satisfaction, and performance of nurses to provide primary healthcare in rural underserved communities. While the components of such a pathway are similar to what have been described earlier for physicians (16), gender-based roles of female nurses and aspirations of government jobs emerged as additional factors to be mindful while planning and evaluating such pathways for PHC nurses. The analysis showed that nurses from rural areas were likely to have a higher retention than those from urban areas. We also noted that these PHC nurses were more likely to have community connect, and performed better. Studies from other contexts have also demonstrated that healthcare staff are more likely to stay on in jobs if they belong to the similar communities they are expected to serve (23,24); which in turn leads to their higher effectiveness (16). At AMRIT Clinics, transition from an ANM or GNM to the role of primary healthcare nurse requires training to build the clinical acumen, understand the context and community needs and to build counseling and management skills. In preparing physicians to undertake rural health, there have been efforts to integrate in their training the values of empathy, social accountability, and community sensitivity (25). We have also observed the value of integrating these values within structured training as well as through mentoring and feedback provided to the PHC nurses. While task-shifting and task redistribution have been used synonymously, we prefer the latter term, as it signifies the principles of equality (as opposed to hierarchy) and of shared responsibility and accountability. Task redistribution in AMRIT Clinics resulted in complementarity between roles of PHC nurses and physicians: while the former played a larger role in managing maternal and child health conditions and communicable diseases, physician played a role in managing non-communicable diseases, providing tele-consultation, as well as in training PHC nurses on weekly visits. Both contributed significantly to the clinic performance. Such an arrangement allows the PHC nurses to perform their role with confidence, and for the team to have high performance. One of the key challenges in providing primary healthcare services in rural and remote areas is sustaining the services through retention and sustenance of healthcare providers. There is no uniform strategy that looks at what works and what does not. Literature on human resource for health (26,27) highlights working conditions and job satisfaction as important for retention. Studies also emphasize on remuneration and career advancement as key strategies as well (8,26). Our experience also elicits the value of these components in building selfconfidence and motivation of the PHC nurses. Additionally, from our experience, it appears that an enabling work environment, simple-to-use technology, and on-going mentoring sustains the motivation of the PHC nurses. The opportunity to undertake responsibilities over and above normative responsibilities within nursing practice is perceived as equivalent to a physician's work. Therefore, the pathway enables the aspirations and potentials of this cadre by also drawing in a valuable role of the physician in shaping it. Studies on nurse practitioners in high-income countries have elucidated how they have been able to provide services equivalent to physicians and have been effective in providing specialized care such as NCDs, maternal health, etc. (8)(9)(10). In our context, we observed that when mandated, skilled, and supported, PHC nurses can meet the complex health needs of rural and tribal populations effectively. In fact, they are likely to promote equitable access to healthcare, by extending care to women and children. Studies on human resource in health highlight several push and pull factors that are responsible for retention of the healthcare workforce in rural areas (26,28). This ranges from working conditions, home and family factors, local environment, and the national and international environment. In our context, the three areas inhibit retention. Firstly, family pressure owing to gendered expectations of the PHC nurses to manage their homes, children, and therefore desired proximity of the workplace closer to their home. Secondly, poor infrastructure of opportunities in the local environment such as housing for spouses, education facilities for children. Thirdly, in the context of India, a pull factor is the desire for a government job due to a higher remuneration offered for permanent nurses and job security. Based on the findings, we propose to actively engage with families to identify ways to resolve the tension between their professional roles and gendered family roles. We also propose to explore public-private-partnerships in primary healthcare, so that some of the disparities in government and private salaries could be resolved, and optimize primary care nurses' availability in rural areas. --- Implications on Policy and Practice In India, rural-urban divide in access to healthcare and health outcomes is related to inequitable distribution of skilled and motivated human resources. The experience and evidence presented here suggests that the rural pathway approach, centered on task redistribution can help in bridging this divide. While recent guidelines in India [as part of Clause 32 of the National Medical Commission Bill (29)] have envisaged the role of nurses (as mid-level providers) in provision of comprehensive primary care in rural areas, it does not provide the pathways to enable them to do so. Our analysis provides some pointers. Sourcing nurses from rural and tribal communities (rather than from urban areas) to serve in rural areas is likely to be more effective since they are more likely to stay and be more empathetic to the communities they serve. A task redistribution approach, where tasks are redistributed between nurses and physicians, who work as a team with a shared responsibility, is likely to be more effective than a task shifting approach, which merely shifts the responsibility to deliver primary health care to nurses alone. While the government has emphasized on higher remuneration to increase the nursing workforce, the current structure of hierarchy within medical practice does not foster a team approach. An emphasis on building non-hierarchical primary healthcare teams that draws on complementary strengths of PHC nurses and physicians will go a long way in advancing rural healthcare. Several forms of support systems, such as close mentoring, ensuring basic amenities, and good living and working conditions would be required to enable PHC nurses to play an effective role in delivering rural healthcare. The rural pathway framework ordinarily focuses on systemic processes that can be influenced to strengthen the workforce. It does not encompass the structural factors, such as gender roles and family and societal expectations from the providers, which appear to significantly affect their availability, retention, and performance in rural areas. Future scoping and research should consider these factors while expanding existing pathways for rural workforce. With dearth of studies in LMICs, the paper provides insights and evidence on a rural pathway approach to enhancing the role of PHC nurses in delivering primary healthcare in rural areas. Such a framework can be used in different contexts to plan and evaluate human resource interventions for improving rural healthcare. --- Limitations of the Study The study is based on experience of one organization in a rural, tribal area and therefore may not be generalizable to all such areas. It does however provide a framework for understanding the factors that would enable nurses in other similar areas to effectively provide primary healthcare. We accepted the reasons that nurses stated for leaving the job, but did not conduct interviews with them to understand in-depth nuances of each case. However, the factors that emerged are helpful to design program level interventions to address retention. Finally, the study did not measure quality of care provided by the nurses. While the large numbers of patient-visits managed by the nurses provide an indication of their acceptance and perceived quality by the patients. --- DATA AVAILABILITY STATEMENT The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. --- ETHICS STATEMENT Ethical approval for this study and written informed consent from the participants of the study were not required in accordance with local legislation and national guidelines. --- AUTHOR CONTRIBUTIONS AA has contributed to the data analysis and writing of manuscript. MD has contributed to the literature review and writing of the manuscript. SB and PM have reviewed and designed the pathway framework as well as edited the manuscript. All authors have contributed to the conceptualization of the paper. --- Conflict of Interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Background: Emerging health needs and uneven distribution of human resources of health have led to poor access to quality healthcare in rural areas. Rural pathways provide an approach to plan and evaluate strategies for ensuring availability, retention, motivation, and performance of human resources for health in rural areas. While effectiveness of primary healthcare (PHC) nurses to deliver primary health care is established, there is not enough evidence on ways to ensure their availability, retention, motivation, and performance. The paper draws on the program experience and evidence from a primary healthcare network (AMRIT Clinics), in which nurses play a central role in delivering primary healthcare in rural tribal areas of Rajasthan, India, to bridge this gap. Methods: Rural, tribal areas of Rajasthan have limited access to functional healthcare facilities, despite having a high burden of diseases. We used the rural pathway approach to describe factors that contributed to the performance of the nurses in AMRIT Clinics. We analyzed information from the human resource information system and health management information system; and supplemented it with semi-structured interviews with nurses, conducted by an independent organization. Results: Most nurses were sourced from rural and tribal communities that the clinics serve; nurses from these communities were likely to have a higher retention than those from urban areas. Sourcing from rural and tribal communities, on-going training in clinical and social skills, a non-hierarchical work environment, and individualized mentoring appear to be responsible for high motivation of the primary healthcare nurses in AMRIT Clinics. Task redistribution with due credentialing, intensive and on-going training, and access to tele-consultation helped in sustaining high performance. However, family expectations to perform gendered roles and pull of government jobs affect their retention.Rural and remote areas with healthcare needs and scarcity of health provisions need to optimize the health workforce by adopting a multi-pronged pathway in its design and implementation. At the same time, there is a need to focus on structural Amin et al. Rural Pathways: An AMRIT Experience factors that affect retention of workforce within the pathway. Our experience highlights a pathway of up-skilling PHC nurses in providing comprehensive primary healthcare in rural and remote communities in Low and Middle-Income Countries (LMICs).

Introduction This article examines the role of social-psychological factors in the development of Ambon Malay in the Netherlands, a heritage language spoken by Moluccan immigrants and their descendants (see the section Ambon Malay in the Netherlands). Due to the intense contact with Dutch, the dominant language of the country, and to the restricted domains of usage (typically the home), heritage Ambon Malay has come to diverge significantly from its homeland variety and has adopted a number of Dutch-like features (Tahitu, 1989;Huwa<unk>, 1992;Aalberse & Moro, 2014;Moro, 2014Moro,, 2015Moro,, 2016Moro,, 2017;;Moro & Klamer, 2015;Moro & Irizarri van Suchtelen, 2017; see also the section Structural features). In the scenario approach to language contact, every linguistic change is seen as both structurally and socially embedded (Muysken, 2010, p. 272). The outcome of language contact is thus determined both by structural and social factors: structural factors, such as the typological profile of the two languages, supply the material for the change and determine the shape that the change is going to take; social factors determine the amount and the destination of the change (Croft, 2000;Muysken, 2010;Johanson, 2013). In heritage language communities, social factors are the main determinants of the variation observed among speakers. Some heritage speakers speak an acrolectal variety closer to the baseline, while others speak a basilectal variety more divergent from the baseline, with a range of possibilities along the continuum (Kondo-Brown 2003, p. 2;Polinsky & Kagan, 2007, p. 371; for a more detailed discussion see section Socialpsychological factors). The goal of this article is to contribute to this line of research by asking: Can socialpsychological factors account for the different frequency of Dutch-like features among heritage speakers? Unlike previous studies, the focus of this article is not on heritage language proficiency, but rather on heritage language structure. More specifically, this article brings together the sphere of language structure, which deals with the structural changes that the heritage variety has undergone, and the sphere of language use and language attitude. Using both a quantitative and qualitative approach, this article investigates whether social-psychological factors predict the frequency of use of certain Dutch-like structures in heritage Ambon Malay. --- Ambon Malay in the Netherlands Ambon Malay is a variety of Malay spoken in the Central Moluccas, in eastern Indonesia. Due to historical events, it is also spoken as heritage language in the Netherlands. After the Republic of Indonesia was proclaimed in 1945, the Dutch government dismantled its Royal Dutch-Indies Army, which was formed mainly by Moluccan soldiers. Political events forced the Dutch government to 'temporarily' bring the soldiers and their families (about 12.500 people) to the Netherlands. After their arrival, the Moluccans were housed in 34 camps spread all over the Netherlands. The camps were situated in rural and isolated areas of the country, where the Moluccans lived segregated until the 1960s. During the first period of residence, little effort was made by the Dutch and by the Moluccans to facilitate the integration process because both parties believed that the stay of the Moluccans in the Netherlands would be temporary In the beginning of the 1960s, it became clear that the stay of the Moluccans would become permanent. Therefore, the Dutch government decided to close the camps and to move the Moluccans to newly built wards on the outskirts of small towns. The social consequence of this housing situation was that Moluccans living in a municipality with a Moluccan ward were less likely to have contact with the Dutch than those living in a municipality without a Moluccan ward (Veenman, 1994). Although the situation is changing, Vermeulen and Penninx (2000, p. 9) report that 'though they [the Moluccans] increasingly disperse from these areas, they are still the least likely of all six groups [Turks, Moroccans, Southern Europeans, Surinamese, Antilleans, and Moluccans] to be found in big cities'. --- Social-psychological factors: age of onset of bilingualism, generation, social network, and attitude In bilingual heritage communities, the emergence and development of linguistic change is usually related to factors such as age of onset of bilingualism (sequential vs. simultaneous) and parental language input (Montrul, 2008;Unsworth, Argyri, Cornips, Hulk, Sorace, & Tsimpli, 2014). Montrul (2008) cites a number of studies showing that simultaneous bilinguals make more errors than sequential bilinguals, and (p. 115) concludes that 'incomplete acquisition appears to be more dramatic in simultaneous than in sequential bilingualism, due to the reduced amount of input received in one language while the language was not yet fully developed'. Parental language input is also an important predictor of linguistic innovations. For instance, Irizarri van Suchtelen (2014) shows that, although all Spanish heritage speakers in the Netherlands diverge considerably from their monolingual peers, heritage speakers with two Spanish speaking parents perform more native-like than heritage speakers who grew up with only one Spanish speaking parent and did not productively use Spanish in childhood. Generation is yet another factor that correlates with language maintenance and language proficiency in heritage language communities (Silva-Corvalán, 1994;Hulsen, 2000). The heritage language has been found to decline with each generation, following the pattern G1 > G2 > G3. Generally speaking, first generation speakers are dominant in the heritage language, second generation speakers have relatively strong skills in both the heritage and the majority languages, whereas third generation speakers are dominant in the majority language; beyond the third generation, few heritage speakers retain a functional command of their language (Benmamoun et al., 2013, p. 79). Social network structure has also proved to be a factor influencing the degree of language preservation and language ability in immigrant communities. Most of the findings demonstrate the territorially based networks enhance the frequency and the intensity of contact among the community members, which, in turn, favour language preservation (Wei, 1994;Wei et al., 2000;Chau, 2011). For instance, a comparison between two Chinese groups of families in Tyneside, England, -one with strong ties and affiliated to the True Jesus church, the other with weak ties and lacking a centralized institution-shows that the families of the first group have a higher level of language maintenance and language proficiency compared to the families in the second group (Wei et al., 2000). Another example is that provided by Chau (2011), who compares Cantonese heritage speakers in Amsterdam (an urban area), where a large community of Cantonese is present, to Cantonese heritage speakers in Venlo (a rural area), who are rather isolated. Heritage speakers in Amsterdam use Cantonese to a larger extent than the Cantonese heritage speakers in Venlo, because the former can share their experiences with their Cantonese speaking peers who live in the neighbourhood, whereas the latter do not have many Chinese peers. According to Chau (2011), the low exposure to Cantonese of Venlo heritage speakers accounts for the many deviations found in their speech (i.e., wrong use of nominal classifiers, inverted double object construction). Other factors are attitude and ethnic identity (Kondo, 1997;Kondo-Brown, 2001;Noels, 2005). The studies of Kondo (1997) and Kondo-Brown (2001) found an association between levels of Japanese proficiency and factors such as Japanese language use at home, degree of family commitment to the heritage language, and attitude. By comparing heritage and nonheritage learners of German in Canada, Noels (2005) shows that heritage students tend to learn German because it is important for their identity and these heritage learners are also more likely to evaluate themselves as skilled in German when compared to learners without a heritage motivation. A serious weakness of these studies, however, is that the analysis is based on speakers' self-ratings of language proficiency (Kondo-Brown 2003, p. 15). Although many studies report a positive interaction between language use and socialpsychological factors, this interaction is more complex than it first appears. Frequent heritage language use can entail a higher rate of maintenance but it can also entail a higher rate of innovations (or accelerated change). Maintaining one's heritage language does not necessarily mean maintaining the homeland-like variety of the heritage language. According to Schmid (2011), speakers who use the L1 and L2 frequently alongside each other may display accelerated signs of linguistic change, because the frequent use of both languages creates the condition for cross-linguistic influence. In her study on German immigrants in Canada and the Netherlands, Schmid (2011, p. 171) concludes that: These findings suggest that change and deterioration of the L1 which may be witnessed among migrant populations may be determined by two opposite poles: speakers who do not use their L1 at all may experience some degree of 'atrophy', while those who live in a bilingual migrant community where L1 and L2 are used frequently alongside each other and mixed to some degree may find themselves sharing in a language with accelerated signs of contact induced change. To conclude, although there is an increasing body of evidence showing that heritage speakers with a history of low exposure to or use of the heritage language exhibit divergent (less homeland-like) patterns, this is not always the case, as both infrequent and frequent use can accelerate on-going language change. --- The study: Social-psychological factors and linguistic divergence in heritage Ambon Malay The aim of this study is to investigate the association between the structural divergence in various areas of heritage Ambon Malay grammar and social-psychological factors in order to account for the variability observed in the heritage speaker population. The data on structural divergence are based on previous work on heritage Ambon Malay (Moro, 2014(Moro,, 2016(Moro,, 2017;;Moro & Klamer, 2015). --- Methodology Two groups of speakers are tested in this study: a group of heritage speakers and a control group of homeland speakers, as laid out in Table 1. All speakers performed the same three tasks: (i) a simultaneous oral description of a set of 14 videos, (ii) an oral description of a set of 68 short video-clips, and (iii) a sociolinguistic interview (for the complete list of videos, video-clips sand the sociolinguistic questions see Appendix 1, 2 and 3 in Moro, 2016). The speakers were coded for five social factors, as illustrated in Table 2: These data were used for the analysis of social factors in the next section. --- Structural features The structures that will be considered here are those referred to as 'Dutch-like' or 'Dutchaligned' features in Moro (2016, p. 221).2 If homeland Ambon Malay has two options, A and B, and option B is also present in Dutch (e.g. a translation equivalent, or the same word order pattern, etc.), heritage speakers tend to use the Dutch-like feature (option B) more often. The Dutch-like features that will be discussed below are: the higher frequency of the pre-verbal marker ada 'EXIST', and of definite marker =nya 'DEF', the increase in the frequency of DO in give-constructions, and of PP and AP in resultative constructions; the higher frequency of the pre-nominal order for the demonstrative itu and the numeral satu 'one'. The pre-verbal marker ada 'EXIST'. Ambon Malay does not mark tense, but it marks grammatical aspect analytically. Aspect marking is optional and the correct temporal interpretation of the utterance is often inferred from the (extra-) linguistic context. In Dutch, present and past tense marking is expressed by means of verbal inflection. Tense inflection and verbal agreement instantiate also the feature of finiteness, which is obligatorily marked in Dutch. The Ambon Malay pre-verbal marker ada mainly marks progressive aspect. the description of the 14 videos, ada is used much more frequently by heritage speakers (13.75%) than by homeland speakers (3.43%). The shift in temporal status and frequency of ada suggests a reanalysis of ada as a present-tense or finiteness marker due to the intense contact with Dutch, a language where tense and finiteness are obligatorily encoded on the verb. The definite marker =nya. Ambon Malay nouns are not obligatorily marked for (in)definiteness. Unlike Ambon Malay, Dutch obligatorily expresses the category of (in)definiteness. The Ambon Malay enclitic =nya may occur on nouns to indicate definiteness. In the description of the 14 videos, =nya is used much more frequently by heritage speakers (11.1% of already mentioned nouns) than by homeland speakers (0.7%). The increase in the frequency of =nya suggests some heritage speakers use this form consistently to mark already mentioned nouns, following a pattern typical of Dutch. The Double Object construction in give-events. 3 To express give-events, Ambon Malay allows both Prepositional Object (PO) constructions (e.g. John gives a book to Mary) and Double Object (DO) construction (e.g. John gave Mary a book). The use of the DO construction, however, is rare and mostly limited to constructions in which the Recipient argument is a pronoun (e.g. John gave her a book). Dutch also allows both constructions, PO constructions are more frequent in elicited data, while DO constructions are very frequent in corpora (Colleman & Bernolet, 2012). Heritage speakers use DO constructions (11.28%) significantly more than homeland speakers (0.74%). The Prepositional Phrase and the Adjectival Phrase constructions in resultative events. Both Ambon Malay and Dutch use various constructions to expression resultative events. Ambon Malay prefers serial verb constructions (SVC) (e.g. She breaks a stick becomes two), while Dutch prefers verb particles (e.g. She cuts off a branch). However, both languages also allow resultative preposition phrases (PP) (e.g. She breaks a stick in two) and adjectival phrases (AP) (e.g. She hits a vase broken). When describing the ten 'cut and break' video-clips, heritage speakers of Ambon Malay use resultative PPs and APs, the constructions shared with Dutch, significantly more than homeland speakers, who prefer SVC. The data are summarized in Figure 2. --- Resultative constructions The pre-nominal order for the demonstrative itu and the numeral satu 'one'. Ambon Malay allows an alternation between two word-order patterns: demonstratives and numerals can precede (DEM-N, NUM-N) or follow the noun (N-DEM, N-NUM). In the description of the 14 videos, homeland speakers have a strong preference for the post-nominal order, both in the case of the distal demonstrative itu 'D.DIST' and the numeral satu 'one'. On the other hand, heritage speakers prefer the pattern shared by Dutch, which only allows the pre-nominal order. In the case of the distal demonstrative itu 'D.DIST', heritage speakers show a neat preference for the DEM-N order (59.17%); in the case of the numeral satu 'one', the trend is less neat but there is an increase in the frequency of NUM-N order (43.24%). The data are summarized in Figure 3. Figure 3. Order of the distal demonstrative itu and the numeral satu 'one' with respect to the noun in homeland Ambon Malay and heritage Ambon Malay speakers. To sum up, heritage speakers tend to prefer the Dutch-like structure over the typical Ambon-Malay structure, even when the former is only marginal in the homeland variety. These changes are on-going, and some speakers are more innovative than others. The following section shows that social-psychological factors can (partially) account for the variation among heritage speakers. --- Analysis: the role of social-psychological factors This section tests the effect of social-psychological factors on the use of Dutch-like structures. Based on previous research, the following six social-psychological factors were tested: onset of Dutch bilingualism, parental language input, generation, place where the speaker grew up, place where the speaker lives, and attitude (as shown in Tables 2 above). Here place where the speaker lives is taken as an indication of social network. Note that some of these factors are correlated: onset of Dutch bilingualism strongly correlates with parental language input (r(32)=.866, p<unk>.001), and also with place where the speaker grew up (r(32)=.673, p<unk>.001). In fact, sequential bilinguals are mostly individuals who grew up in households in Moluccan camps, where Ambon Malay was the main language spoken. Due to this correlation, the factors parental language input and place where the speaker grew up were excluded from the analysis. Following common practices in variationist sociolinguistics, the effects of socialpsychological factors on speakers' performance were assessed using a multivariate general linear model (in SPSS). Given the small sample size (only 32 speakers), we can only expect to find statistical significance for the strongest effects, and where we do not find a significant effect it may either be that there is no such effect or that it is too weak to surface in the current sample. The results of the general linear model analysis are summarized in Table 3 and discussed below. b. R Squared =.777 (Adjusted R Squared =.586). c. R Squared =.790 (Adjusted R Squared =.610). d. R Squared =.591 (Adjusted R Squared =.240). e. R Squared =.556 (Adjusted R Squared =.176). f. R Squared =.588 (Adjusted R Squared =.236). g. R Squared =.431 (Adjusted R Squared = -.057) The general linear model analysis indicates that generation is not a predictor of Dutch-like feature use, as it does not play a role for any of the features. The place where the speaker lives has a significant effect for the features =nya 'DEF' (p =.002), DO constructions (p =.006), and pre-nominal itu 'D.DIST' (p =.007), such that heritage speakers living outside a Moluccan ward use these Dutch-like features more frequently. Onset of Dutch bilingualism has a significant effect for =nya 'DEF' (p =.008) and DO constructions (p =.026), such that sequential bilinguals use these Dutch-like features more frequently. Attitude has a significant effect only for DO constructions (p =.033), such that speakers with a mild positive attitude use this Dutch-like feature more often than speakers with a strongly positive attitude. The interaction between place where the speaker lives and onset of Dutch bilingualism has a significant effect for the features =nya 'DEF' (p <unk>.001) and DO constructions (p =.002), and a trend toward significance for PP constructions (p =.051), such that sequential bilinguals living outside a Moluccan ward have the highest rate of these three Dutch-like features. The interaction between place where the speaker lives and generation also has a significant effect for the features =nya 'DEF' (p =.030) and DO constructions (p =.048), and a trend toward significance also for pre-nominal itu 'D.DIST' (p =.090), such that speakers of the second generation and of the generation 2.5 living outside a Moluccan ward are the most innovative. Finally, the interaction of onset of Dutch bilingualism and attitude has a significant effect for the use of AP constructions (p =.002), and a trend toward significance for PP constructions (p =.090), such that simultaneous bilinguals with only a mild positive attitude are the most innovative. The general linear model has shown that the place where the speaker lives is the best predictor of Dutch-like features, followed by onset of Dutch bilingualism and attitude. A Bonferroni posthoc test was performed after any significant main effect. The results are summarized in Table 4 and discussed below. Speakers living outside a Moluccan ward have a higher rate of Dutch-like features (except for AP and PP constructions) than speakers living inside a Moluccan ward. Similarly, speakers with only a mild positive attitude toward the heritage language have a higher rate of Dutch-like features (except for AP constructions) than speakers with a strongly positive attitude. As for onset of Dutch bilingualism, the trend is less clear; in the case of =nya 'DEF', DO and AP constructions, sequential bilinguals are more innovative, while for the remaining Dutch-like features, it is simultaneous bilinguals who are more innovative. Three case studies. This section describes three case studies of Ambon Malay heritage speakers, discussing their socio-psychological profile and their linguistic performance with respect to the above-mentioned Dutch-like features, see the comparison in Table 5 below. The first speaker is Max (a fictitious name), a 45 years old man who belongs to the 2.5 generation. He grew up as simultaneous bilingual in Amsterdam, where he still lives today. He has a positive attitude toward the heritage language but he is not committed to heritage language maintenance and has not transmitted it to his daughter. He married a Dutch woman of non-Moluccan origin. He reports having both Dutch and Moluccan friends, and to occasionally speak heritage Ambon Malay with some of them. The heritage Ambon Malay he speaks presents a high rate of all the Dutch-like features (except for AP constructions). The second speaker is Tom (a fictitious name), a 59 years old man who belongs to the second generation. He grew up as sequential bilingual in a Moluccan camp, but nowadays he lives in a city (outside any Moluccan ward). He is married to a Dutch woman of non-Moluccan origin. His attitude toward Ambon Malay is positive, but he thinks that the heritage language is not important, it becomes important only if one wants to move back to the Moluccas. The heritage Ambon Malay he speaks presents a high rate of all the Dutch-like features (except for pre-nominal satu 'one'. The third speaker is Jan (a fictitious name), a 27 years old man who belongs to the third generation. He grew up as simultaneous bilingual in a Moluccan ward, where he still lives today. He has a very positive attitude toward the heritage language; he married a Dutch woman of Moluccan origin who also grew up in a Moluccan ward, and they are transmitting the heritage language to their daughter. He speaks Ambon Malay to his wife, and to his grandparents, who raised him. He reports feeling proud to be able to speak Ambon Malay. The heritage Ambon Malay he speaks is very homeland-like, with a low rate or absence of the above-mentioned Dutch-like features. To conclude, the case studies show that generation does not play a role, as the speaker of the third generation, Jan, is the most conservative. Onset of Dutch bilingualism does not have a strong influence, because Max, a simultaneous bilingual, and Tom, a sequential bilingual, are both innovative. The place where the speaker lives and attitude seem to make a difference, Jan, who lives in a Moluccan ward and has strongly positive attitude toward the heritage language, speaks the variety less influenced by Dutch. --- Discussion The previous section tested the association between social-psychological factors and frequency of Dutch-like features to answer the question: Can social-psychological factors account for the different frequency of Dutch-like structures among heritage speakers? The factors place where the speaker lives, onset of Dutch bilingualism, and attitude all have a significant effect on the frequency of Dutch-like features among heritage Ambon Malay speakers. The place where the speaker lives and attitude have the expected effect, namely speakers living inside a Moluccan ward and with a strongly positive attitude are more conservative, while speakers living in a city (outside a Moluccan ward) and with only a mild positive attitude are the most innovative. Contrarily, onset of Dutch bilingualism does not have a clear effect. For some of the variables, it is sequential bilinguals who are more innovative, while following the literature, we would expect them to be more conservative, as they have been exposed solely to Ambon Malay during the first years of childhood. The first issue that this section addresses is, then, why onset of Dutch bilingualism does not have a straightforward effect in the shaping of heritage Ambon Malay grammar. We have seen that, in some cases, sequential bilinguals show a higher rate of innovation than simultaneous bilinguals, and that second generation speakers seem more innovative than third generation ones. This apparently counterintuitive finding is actually in line with the observation of Schmid (2011), namely that language change may be found among speakers who frequently use their L1 and L2 alongside each other. Now, it is plausible to assume that, having received considerable Ambon Malay input in childhood, sequential bilinguals and second generation speakers are more confident about their linguistic skills in the heritage language, and thus use the language more frequently than the other speakers. The frequent use of the dominant and the heritage languages alongside each other creates the conditions for cross-linguistic influence. Thus, if high exposure in childhood translates into relatively frequent use in adulthood, then the language of sequential bilinguals can indeed show 'accelerated signs of contact induced change'. Another issue related to onset of Dutch bilingualism is the nature of the linguistic variables that are tested. The innovative features described in this study are rather different from the features described by Montrul (2008) and Unsworth et al. (2014). The phenomena investigated by these authors are errors in tense-aspect-mood inflection morphology or in the assignment of grammatical gender. For this type of phenomena, age of onset of bilingualism has proved to be an important factor predicting the incomplete acquisition of morphological marking (but cf. Kupisch, 2013). The phenomena described in this study are not errors in inflectional morphology, but rather changes in frequency between structures already present in Ambon Malay. Acquiring the form-meaning mapping of a particular construction, together with its frequency, pragmatics, and contexts of usage may be a process that continues well beyond the age of five. Hence, onset of Dutch bilingualism may not make a difference, because the innovative features of Ambon Malay are not acquired once and for all in childhood but are part of a gradient process of language acquisition that continues through the lifetime and is sustained by other intervening factors, such as the amount and the type of Ambon Malay speaking contacts an individual has. This observation supports the importance of social network, to which we now turn. Social network, operationalized here as the place where the speaker lives, is by far the most important factor in the heritage Ambon Malay community. Heritage speakers living outside a Moluccan ward show the highest rate of Dutch-like features, whereas speakers living in a Moluccan ward retain homeland Malay-like features more firmly. This finding agrees well with previous studies on the immigrant Chinese community (Wei, 1994;Chau, 2011) in the U.K. and other studies on the Ambon Malay community (Huwa<unk>, 1992;Veenman, 1994;Tahitu & Lasomer, 2001), which found an association between social network and language proficiency and language maintenance. The most straightforward explanation for this association is that heritage speakers living outside a Moluccan ward have fewer chances to speak Ambon Malay compared to their peers living inside a Moluccas ward, so their heritage language shows signs of 'atrophy'. Since they are fully immersed in a Dutch-speaking environment, there are higher chances that, when they speak Ambon Malay, they will rely on Dutch-like structures. Another reason why heritage speakers living in a city (outside a Moluccan ward) are more innovative is that they are more likely to be part of a social network of weak ties (Milroy & Milroy, 1985). Thanks to their'mobility' and the many acquaintances they have (weak ties), these individuals are more likely to be exposed to new (linguistic) information and to pass them on. The process of propagation of linguistic innovations has been described by Enfield (2003, p. 366) in the following way: Some individual or individuals begin to habitually perform a new linguistic act, exposing those in their personal network to the idea, with the result that those who are exposed then replicate this performance (given sufficient motivation to do so), and in turn expose more people in their own social networks (as well as those who began the process in the first place, revalidating and encouraging the usage, and leading it to take further hold). Thus, it is likely that Ambon Malay heritage speakers living in a city participate in several networks and interact with (at least some) individuals in Ambon Malay; when they move from one network to another, they may (more or less voluntarily) diffuse the innovations that they have picked up in these former interactions. Their language thus shows signs of accelerated change. The interaction between place where the speaker lives and onset of Dutch bilingualism indicates that, among the speakers living in a city, the most innovative are sequential bilinguals, who supposedly use their heritage language frequently. This finding agrees with the observation that in order to be an innovator, a speaker needs to use his/her (heritage) language (Schmid, 2011). As mentioned above, sequential bilinguals probably feel more comfortable speaking the heritage language than (some) simultaneous bilinguals and this leads them to use Ambon Malay and Dutch alongside each other frequently. The continuous and frequent switch from one language to the other is likely to increase the chances of cross-linguistic influence, eventually leading to a shift in favour of Dutch-like structures. Finally, attitude also influences the frequency of Dutch-like features. Heritage speakers with a strongly positive attitude retain homeland Malay-like features, while heritage speakers with only a mild positive attitude use Dutch-like features more frequently. This finding is in line with those of previous studies reporting an association between attitude and language proficiency (Kondo, 1997;Noels, 2005). In my sample, all heritage speakers with a strongly positive attitude have transmitted the heritage language to their children or are planning to do so, all (but one) have a spouse of Moluccan origin, some of them have joint Moluccan political or cultural organizations, one visits the Moluccas almost every year, another one writes songs in Ambon Malay and has participated to a workshop on language revitalization. In these settings, heritage speakers are likely to interact with monolingual speakers of Ambon Malay, or with other heritage speakers but without code-switching. In other words, these setting are conducive to maintain a baseline variety of the heritage language, one without much Dutch influence. --- Conclusion Heritage Ambon Malay in the Netherlands has been found to diverge from the homeland variety spoken on Ambon, Indonesia, and to converge toward Dutch. Convergence is mainly instantiated by changes in frequency between two structures, in favour of the structure which is shared with Dutch (Dutch-like feature). This study investigated what social-psychological factors predict the use of Dutch-like features among heritage speakers of Ambon Malay. The place where the speaker lives, taken here as an indication of social network, strongly predicts the amount of innovation in the heritage language: individuals living outside Moluccan wards are the ground breakers in most of the changes. Onset of Dutch bilingualism does not always have the expected effect; in some cases, sequential bilinguals are subject to more cross-linguistic influence (higher rate of Dutch-like feature) than simultaneous bilinguals, because they frequently use the two languages alongside each other. Finally, speakers with a very positive attitude will create and exploit all opportunities they have to speak the heritage language, which has a beneficial effect on the maintenance of homeland Malay-like features in their heritage language.

Aims and Objectives/Purpose/Research Questions This article examines the role of social-psychological factors in the development of heritage Ambon Malay in the Netherlands. More specifically, it aims to answer the question: Can socialpsychological factors account for the different frequency of Dutch-like structures among heritage speakers? Design/Methodology/Approach Data from 32 Ambon Malay heritage speakers and 27 Ambon Malay homeland speakers were collected by means of video stimuli and a sociolinguistic interview.The database provides six linguistic variables and three social-psychological factors. The linguistic variables are: the pre-verbal marker ada, the definite marker =nya, the double object construction, the prepositional phrase and adjectival phrase in resultative constructions, the prenominal order for the demonstrative itu and the numeral satu 'one'. The social-psychological factors are place where the speaker lives, onset of Dutch bilingualism, and attitude. The effect of the social-psychological factors on the linguistic variables was assessed using a multivariate general linear model.The results show that place where the speaker lives is the best predictor. Heritage speakers living outside a Moluccan ward have a higher rate of Dutch-like features than speakers living inside a Moluccan ward. In some cases, sequential bilinguals are more innovative than simultaneous bilinguals. Finally, speakers with only a mild positive attitude toward the heritage language have a higher rate of Dutch-like features. Originality Unlike previous studies, this article does not test the role of social-psychological factors against self-ratings of heritage language proficiency, but it uses real language data.The theoretical significance of this research is to bridge the gap between the sphere of language structure and the sphere of language use and language attitude. An additional value is in its finding that frequent use of the heritage language means higher rate of maintenance but also accelerated change.

Introduction Working From Home (WFH), which refers to work performed at a remote location (such as home) [1], is not a novel concept. The development of information technology and a multitude of internet-based platforms has made team collaboration and distant communication between leaders and employees and between employees and their clients increasingly convenient. The benefits of WFH may include increased job satisfaction, reduced travel time and expenses, increased productivity, and reduced turnover and absenteeism [2][3][4][5]. The drawbacks of WFH may include isolation from the work culture, potential conflicts between work and home, a lack of control over employees, difficulties in teamwork, and so on [2]. A meta-analysis [3] demonstrated that WFH had small but mainly beneficial effects on proximal outcomes, such as perceived autonomy and (decreased) work-home conflict. WFH also had beneficial effects on more distal outcomes, such as job satisfaction, performance, turnover intention, and role stress. Hence, it is unclear whether WFH affects employee wellbeing positively or negatively, as the evidence from the existing literature is indeterminate and often contradictory. Since the outbreak of the COVID-19 pandemic, the concept of WFH has gained increased attention and additional meaning, as the proportion of people who worked from home experienced an unprecedented increase during the pandemic. For instance, Bick and colleagues [6] found that 35.2 percent of the workforce in the United States worked entirely from home in May 2020, up from 8.2 percent in February 2020. In addition, for many employees, the change in the overall work situation has been sudden and mandatory, leaving them unprepared for the transition [7]. Thus, the existing literature on WFH may not be applicable to the pandemic situation [8]. Our study aimed to contribute to the literature on WFH during the COVID-19 pandemic in the following three ways. First, as previous studies have conducted some descriptive analyses, such as the demographic characteristics of WFH employees after the outbreak of the pandemic [6,9], more empirical studies are needed to further examine how theoretical models such as the Job Demands-Resources (JD-R) model works in the WFH context during the COVID-19 pandemic. Second, although previous studies compared employees WFH and employees going to work, accurate comparisons were difficult, and the results are inconsistent, as groups of employees WFH are usually much smaller than those of in-office employees [1]. Thus, lockdown policies, including the transition to WFH during the COVID-19 pandemic, offer a unique opportunity to explore differences in experience between employees who worked from home and those who still went to work. Third, different countries responded differently to the COVID-19 pandemic, and few studies have contributed to the literature on WFH during the pandemic in a Norwegian context. In Norway, a national lockdown was announced on 12 March 2020 [10], and since then, working from home has been the main policy of many organizations. Nordic countries, such as Norway, have dealt with the COVID-19 crisis more efficiently than most other countries [11] due to the ease with which citizens maintain social distance and thus prevent spreading the virus. Norway also has a public welfare system that is well adapted to reduce the negative influence of redundancies, unemployment insurance, and sick leave. Therefore, in this study, based on data collected in Norway in February 2021, we drew from the Job Demand-Resources model [12,13] as well as the Conservation of Resources (COR, [14,15]) theory and aimed to explore how relationships between job demands (workhome conflict and workload), job resources (coworker support, leader support, and familywork facilitation), and wellbeing (burnout and work engagement) have been experienced by employees who changed from working in the workplace before the pandemic to working from home during the pandemic (change group) and those who remained working in their workplaces before and during the COVID-19 pandemic (no-change group). --- The COVID-19 Pandemic, Work Situation, and Wellbeing The pandemic has been proved to increase mental health risks to a large extent. For instance, a longitudinal study from the UK reported that the percentage of individuals with an anxiety disorder almost doubled during COVID-19 at 24% (95% CI 23%, 26%) compared to the pre-pandemic level of 13% (95% CI 12%, 14%) [16]. Research has also illustrated an everyday life characterized by loneliness and poorer mental health in the COVID-19 context [17]. In Germany, researchers have demonstrated that the lockdown circumstances generally have a negative effect on the satisfaction with the work and home life of individuals [18]. Some researchers have already made great efforts to investigate the influence of WFH during the pandemic on employees' mental health and work-related attitudes or behaviors. According to Carillo et al. [7], epidemic-induced WFH inherits some of the characteristics from conventional WFH, but it appears to have particular aspects that make it a unique context with specific conceptual boundaries. First, it became more of a mandatory requirement than a voluntary option. Individuals across occupations who preferred going to work were forced into WFH [19]. Second, it occurred not only in certain businesses that may have a preference for WFH and whose employees are more productive if they are WFH, such as internet companies, but also in other traditional occupations (e.g., teachers in education). Thus, employees all over the world have worked substantially more from home, with their social skills in greater demand, being forced to communicate with their colleagues and clients remotely due to restrictions [20]. Further, employees may face challenges due to fundamental issues such as lacking space in one's home to attend to work and having difficulty receiving instant or efficient feedback from students or colleagues [21]. Research on WFH during the pandemic has revealed mixed results. Employees who worked from home had a high sense of insecurity, increased work-home conflict, and high levels of stress [22,23]. This was further supported by Hayes et al. [24], who indicated a higher level of burnout and stress among US employees WFH compared to before the pandemic when they went to the workplace. It was also reported that WFH was associated with higher job performance due to its negative association with stress in a Canadian population [23]. Moreover, Dubey and Tripathi [25] analyzed Twitter activity and found that the WFH concept was viewed positively by employees. Thus, researchers and practitioners are concerned with the influence of the workplace change on employee wellbeing during the COVID-19 pandemic [24]. Meanwhile, for employees who remained working on the frontline under the pandemic, such as health care workers, virus-related stressors such as the fear of being infected as well as the fear of infecting their family have also become major concerns, which may be detrimental to their wellbeing [26][27][28]. Despite these findings, few studies have investigated both employees who worked from home and those who remained in the workplace. In the pandemic context, a qualitative study [8] identified four key remote work challenges (work-home interference, ineffective communication, procrastination, and loneliness), as well as four virtual work characteristics that affected the experience of these challenges (social support, job autonomy, monitoring, and workload). Therefore, work-home conflict and workload can be identified as important demands that are highly related to WFH. In addition, social support appears to be a beneficial resource for employees under a lockdown situation when they may face the challenge of loneliness. In particular, employees may have different perceptions about different sources of social support (e.g., leaders, coworkers, and family support) under such circumstances. Furthermore, burnout and engagement have been considered important indicators of employee wellbeing and have been shown to be highly related to work-related attitudes and behaviors such as job satisfaction, commitment, and performance [29,30]. In our study, we aimed to explore the importance of workload, work-home conflict (WHC), and social support, including leader support, coworker support, and family-work facilitation (FWF), in relation to wellbeing (burnout and work engagement) among employees who still went to the workplace (no-change group) and employees who transitioned to WFH (change group) during the COVID-19 pandemic. At the time of the data collection, Norway was working to curb the second wave of infections and was under lockdown to prevent overloading the healthcare system. Thus, employees in these two groups may have experienced their work situations differently during the period of study. Considering the previous literature indicating that employees in both the change and no-change group may encounter different stressors under the pandemic, we propose the following research question: Research question 1 (RQ1). Are there significant differences in levels of WHC, workload, coworker support, leader support, and FWF between employees in the change and no-change group? --- Theoretical Framework Given the variables that we included in our study, we combined the JD-R model and the COR theory to explain our study hypothesis. According to the JD-R model, all types of job characteristics can be classified into one of two categories: job demands and job resources [12,13]. Job demands are defined as physical, psychological, social, or organizational aspects of the job that require sustained physical and/or psychological effort and are therefore associated with certain physiological and/or psychological costs [13]. On the other hand, job resources refer to physical, psychological, social, or organizational aspects of the job that are necessary to achieve work goals, reduce job demands and the associated physiological and psychological costs, or stimulate personal growth, learning, and development [13]. Burnout is traditionally characterized as a syndrome of exhaustion, cynicism, and lack of efficacy experienced by employees [31][32][33]. In contrast to burnout, engagement is a positive, fulfilling, work-related state of mind characterized by vigor, dedication, and ab-sorption [34]. Numerous researchers have reported that job demands are unique predictors of exhaustion, whereas job resources are unique predictors of engagement [29,[34][35][36][37]. In addition, in this study, the COR theory may serve to explain the importance of resources outside of work, such as family-work facilitation. According to this theory, humans are driven to foster, create, preserve, and protect their resources [38]. A central principle in the theory is the gain spiral, where individuals with greater resources find it easier to obtain new resources. In addition, people who possess resources do encounter stressful situations, but they are better equipped to deal with these stressors [38]. --- Relationship between Job Demands/Job Resources and Burnout/Work Engagement in the Context of WFH As is widely supported by the literature and the JD-R model, WHC has been an important issue related to WFH [3]. As the current pandemic continues to unfold, the potential for conflict between work and home spheres may be greater than ever [39]. Some researchers have focused on the effect of WHC on productivity during the COVID-19 pandemic [40], but more empirical studies are still needed to further investigate the influence of WHC during the pandemic from different perspectives. For instance, comparing the relationship between WHC and employee wellbeing among the change and no-change group may better reflect the overall influence of WHC during the pandemic. Moreover, researchers have revealed that workload has been a noteworthy job demand during the pandemic, as some employees stated that their work seemed never-ending when working from home [8]. Some employees also believed that people could decide to work right away or procrastinate while working from home, which was associated with their workload. Sadiq [41] found that WHC mediated the relationships of workload with job stress and job dissatisfaction in Pakistani police. Based on the above, we consider WHC and workload to be important job demands that may influence employees' burnout and work engagement. In sum, we intended to test the following hypotheses among both the change and no-change group: Hypothesis 1 (H1). WHC is (a) positively related to burnout and (b) negatively related to work engagement. --- Hypothesis 2 (H2). Workload is (a) positively related to burnout and (b) negatively related to work engagement. When it comes to job resources, researchers have largely determined that social support plays a beneficial role for employees who work from home. That is, if employees receive more social support, they will perform better when working from home. For example, Wang et al. [8] utilized quantitative data to prove that social support had a positive effect on performance due to less procrastination and home-work interference; a negative effect on emotional exhaustion due to less procrastination, work-home interference, and loneliness; and a positive effect on life satisfaction due to less work-home interference and loneliness. Szkody et al. [42] also indicated that social support buffers the connection between worry about COVID-19 and psychological health. Moreover, Xiao et al. [43] proved that social support provided to medical staff caused a reduction in anxiety and stress levels and increased their self-efficacy. Data from Thailand also showed that supervisor support had a negative effect on employees' emotional exhaustion [44]. However, the beneficial effect of social support has been challenged, as Deelstra et al. [45] proved that the reception of support evokes feelings of incapacity, which in turn evokes poor self-image in employees. Furthermore, the literature highlights that during the COVID-19 pandemic, social support has changed to digital support, and consequently, it is unclear how effective support is in improving employees' state of wellbeing [8,39]. In this study, we included three different sources of social support, namely, leader support and coworker support as job resources and FWF (family support) as a home resource, to comprehensively examine the effects of different kinds of social support on burnout and work engagement. In line with the COR theory, workers who experience high FWF will be in a better position to invest these resources in their behavior and attitudes to deal with demanding situations and foster other resources. For example, workers who see their family as benefiting from their work will consequently be more energized and enthusiastic toward their work [46] and, thus, be more engaged. Furthermore, we are interested in how these three sources of social support differ in relation to employee wellbeing to see which is more important for employees in the change and no-change group. Therefore, we propose the following hypotheses for both the change and no-change group: --- Materials and Methods --- Study Design The present study was based on a larger student project named "Healthy workplaces in light of COVID-19" using a questionnaire of previously validated scales related to wellbeing at work. The online survey was conducted using the survey service "Nettskjema", provided by the University of Oslo in Norway. The invitation, consent information, and the link to participate in the study were distributed by undergraduate students via email. Employees working in (1) scientific, technical, and administrative services and (2) health and social services were invited to participate in the study. These occupational groups were selected to collect data from employees most likely to work from home during lockdown (scientific, technical, and administrative services) and from employees likely to go to work during the pandemic (health and social services). Convenience (acquaintances) and snowball sampling methods were implemented between 25 January 2021 and 7 February 2021. This was during a period where the level of lockdown in Norway would indicate that most employees were working from home if they had the opportunity to do so. The study was conducted in accordance with the guidelines of the Norwegian Center for Research Data (NSD). --- Instruments All the scales were measured on a 5-point Likert scale, with responses ranging from never to always (N-A) or from totally disagree to totally agree (D-A). In addition to the variables applied in this study, the survey also included demographic measures such as gender and age. Utrecht Work Engagement Scale-3 (UWES, [47]) is a shortened version measuring the employee experience of being engaged at work with 3 items (a = 0.78). Each of the 3 items was meant to capture one of the subdimensions of work engagement-absorption, vigor, and dedication-on a scale ranging from N to A. One example is, "At my work, I feel bursting with energy". Burnout Assessment Tool (BAT, [48,49]) measures the employee experience of being burned out with 12 items (a = 0.88). BAT is composed of 4 subdimensions, including exhaustion, mental distance, and emotional and cognitive impairment, each measured with 3 items. All the items ranged from N to A. An example is: "At work, I feel mentally exhausted". Family-to-Work Facilitation (FWF, [50]) measures the level of positive spillover from family to work and consists of 3 items (a = 0.77) scored from D to A. One example is, "My family helps me acquire skills and this helps me be a better worker". Work-Home Conflict (WHC, [51,52]) measures the level of negative spillover from work to family and consists of 4 items (a = 0.84) ranging from D to A. One example is, "Stress at work makes me irritable at home". The last three scales were obtained from "Questionnaire on the Experience and Evaluation of Works" [53] and were all scored from N to A: (1) Social Support Coworker (SSC) measures social support from coworkers with 3 items (a = 0.72), such as: Can you count on your colleagues for help and support, when needed?" (2) Social Support Leader (SSL) measures social support from managers with 3 items (a = 0.87), such as: "Do you feel your work is recognized and appreciated by your supervisor?" Finally, ( 3 --- Sample In total, 629 participants completed the survey. Out of the total sample, 575 participants met the criteria for inclusion in this study as either belonging to the change or no-change group. Three hundred and eighty-three of these participants were women (66.6%), and most of the respondents were in the age groups <unk>31 (27.8%), 41-50 (19.8%), and 51-60 (32.7%). The sample consisted of employees representing health and social services (43.5%), professional, scientific, and technical services (48.0%), and other (8.6%). A large proportion of the participants reported having completed a higher education of 3 years or more (78.8%). Three hundred and ten participants were living with children (53.9%), and four hundred and four lived with a partner (70.3%). The no-change group consisted of 269 participants who went to work every day both pre-COVID-19 and during the COVID-19 pandemic; hence, their work situation did not change due to the pandemic. The change group was composed of 306 participants who changed their work situation from going to workplaces before the pandemic to working more from home during the pandemic (see Table 1). In the no-change group, 16 percent were employed in professional, scientific, and technical services, and 75 percent were in health and social services. Conversely, in the change group, 76 percent were employed in professional, scientific, and technical services, and 16 percent were in health and social services. --- Data Analysis Descriptive analysis was conducted in IBM SPSS Statistics for Macintosh, version 27 (Armonk, NY, USA), to determine the central tendency and dispersion of the sample, the mean and standard deviation for continuous variables, and frequencies for categorical measures. In addition, the Pearson product-moment correlation coefficient (Pearson's r) was used to measure the bivariate correlations between study variables. The group means of the study variables were then compared through 7 independent sample T-tests with false discovery rate (FDR) 0.05 p value cut-off correction. FDR is a procedure that provides the opportunity to control for the probability of Type I errors while, unlike the Bonferroni correction, allowing a reduction in Type II error [54]. The relationships between the predictors and the outcomes were investigated by multivariate structural equation modeling (SEM) in MPLUS version 8.4 [55]. As the model test statistics of SEM models (<unk> 2 ) are sensitive to sample size, it is recommended to also consider the Standardized Root Mean Square Residual (SRMR) and three approximate fit indices, namely, Steiger-Lind Root Mean Square Error of Approximation (RMSEA), the Bentler Comparative Fit Index (CFI), and the Tucker-Lewis Index (TLI), to evaluate the model's goodness of fit [56]. According to Hooper et al. [57], CFI and TLI values >0.90 and SRMR and RMSEA values <unk> 0.08 indicate an acceptable fit. As all of the variables applied in this study were measured using a self-report questionnaire, we followed the recommendations of Podsakoff et al. [58] to address limitations due to common method bias, evaluating the goodness of fit of the one-factor model compared to that of the study model. --- Results --- Assumptions for T-Test and Structural Equation Model The assumptions of normality and homogeneity were considered before conducting the T-test. Skewness and kurtosis levels were below the recommended threshold of 1 for all the variables in both groups. Levene's test was significant for one variable; thus, the corrected test results were reported for leader support. Before conducting the SEM analysis, we tested for the violation of the assumptions of linearity and multicollinearity. The curve estimation for all the relationships in both groups was sufficient for covariance-based SEM. Additionally, the multicollinearity statistics indicated that the assumption was not violated, mean VIF = 1.298. --- Descriptive Statistics and Independent Sample T-Test The bivariate correlations between the study variables for the two groups (change and no-change) are displayed in Table 2. The strongest correlation was between burnout and WHC in the no-change group and between coworker and leader support in the change group. The descriptive statistics of the study variables are displayed in Table 3. The highest mean value was found for coworker support in the no-change group, whereas work engagement presented the highest mean value in the change group. In both the no-change and change group, the lowest mean score was found for burnout. The results of the independent sample T-tests evaluating the mean differences in the study variables between the no-change and change group are also presented in Table 3. The no-change group (M = 4.03, SD = 0.58) experienced significantly more coworker support (t (573) = 2.534, p <unk> 0.05) than the change group (M = 3.90, SD = 0.64). In addition, the change group (M = 3.53, SD = 0.60) experienced a higher workload (t (573) = -2.989, p <unk> 0.05) than the no-change group (M = 3.37, SD = 0.65). No significant results were found for the other variables. --- Test for CFA and Common Method Bias The model fit indices for the CFA of the study model indicated an acceptable model fit when the four subdimensions of BAT were included (N = 575, <unk> 2 (439) = 1005.22, p <unk> 0.05; CFI = 0.920, TLI = 0.910, SRMR = 0.053, RMSEA = 0.047, 90% CI [0.044, 0.051]). When testing for common method bias, the one-factor model did not provide acceptable fit indices (N = 575, <unk> 2 (418) = 1652.90, p <unk> 0.05; CFI = 0.825, TLI = 0.793, SRMR = 0.435, RMSEA = 0.072, 90% CI [0.068, 0.075]), and hence, common method bias did not seem to be an issue. The control variables, age and gender, were not included in the measurement model. --- The Structural Equation Model The results of the analysis of the SEM model for work engagement and burnout in the change and no-change group are presented in Table 4 as well as Figure 1 The results provide partial support for hypothesis 2 in both the no-change and change group; workload was positively related to (a) burnout only in the no-change group (b = 0.14, p <unk> 0.05), whereas for the change group, workload was positively, not negatively, related to (b) work engagement (b = 0.20, p <unk> 0.01). Testing hypotheses 3, 4, and 5 regarding job resources, leader support, coworker support, and FWF also provided mixed results. Hypothesis 3 was partly supported, as leader support was negatively related to (a) burnout only in the change group (b = -0.17, mboxemphp <unk> 0.05). Further, hypothesis 4 was also partly supported for both groups, as coworker support was positively related to (b) work engagement in both the change (b = 0.20, p <unk> 0.05) and no-change group (b = 0.19, p <unk> 0.01). Lastly, the results provided support for hypothesis 5a, as FWF was negatively related to burnout only in the change group (b = -0.15, p <unk> 0.01). Hypothesis 5b was supported in both groups, as FWF was positively related to work engagement in the no-change (b = 0.15, p <unk> 0.05) and change group (b = 0.26, p <unk> 0.001). --- Discussion The aim of this study was two-fold: (1) to explore differences in the experience of job demands and resources between employees who still went to their workplaces (no-change group) and employees who transitioned to WFH (change group) during the pandemic; (2) to explore the relationships among job demands, job resources, and wellbeing outcomes for employees in these two groups. Our results indicate that the work situation may have deteriorated somewhat for employees in the change group, as they experienced less support from their coworkers and a higher workload than employees in the no-change group (RQ1). There appeared to be no significant difference in other demands and resources. One explanation for this result is that at the time of data collection in this study, employees in Norway had been experiencing lockdown policies for almost one year. As a result, employees in the change group may have adapted to this "new normal". The other reason may be due to the unique context in Norway. Due to protection from the Norwegian Working Environment Act, employees may suffer less from employment issues caused by the pandemic, such as unemployment, sick leave, and health risks. Employers are obligated to ensure that work does not negatively affect employees' physical and mental health. That is, despite the shock of the COVID-19 pandemic, employees can receive support from their employers to ensure that they have fewer job demands and more job resources, which can prevent the development of negative work-related outcomes. The third reason could be that employees in the change and no-change group may experience distinct stressors. As we mentioned before, employees in the change group may face the possible challenges of WFH while employees in the no-change group may encounter the fear of infection as they still must keep close contact with the public. Employees in the two groups may all have experienced work stress during the pandemic. Moreover, we argued that it was necessary to further test the relationship between job demands and resources and burnout and engagement in a Nordic context during the COVID-19 pandemic. To the best of our knowledge, this study is the first to test these relationships separately for employees in the change and no-change group. Regarding job demands, the relationship between WHC (H1) and wellbeing was mostly detrimental in both groups, increasing burnout and reducing work engagement. In addition, workload (H2) was associated with increased burnout solely in the no-change group but was positively related to more work engagement in the change group. When looking at job resources, leader support (H3) was an important factor and was negatively related to burnout in the change group, whereas coworker support (H4) was positively related to work engagement in both groups. Lastly, FWF (H5) seemed to be the most important source of support for the change group, as it appeared to relate to both burnout and work engagement in a positive way. FWF was also positively associated with work engagement in the no-change group. We further summarize the results of our study in the following points: First, our study reflects the JD-R model on the relationships between job demands and burnout as well as job demands and work engagement. Based on previous literature [8], we believed that WHC and workload were two noteworthy job demands for employees who worked from home during the COVID-19 pandemic. Our results support previous studies [26,39,59] that indicate that WHC could be the most influential factor related to burnout and work engagement, with relatively high coefficients in both the change and no-change group. As for workload, for employees in the change group, a greater workload was not related to more burnout but was related to more work engagement, which was a surprising finding in this study. For employees in the no-change group, a higher workload was positively associated with more burnout, as expected. One reason that may explain the unexpected finding is that workload could be considered a challenging demand [60][61][62] for employees who work more from home. A higher workload can motivate and encourage employees to concentrate on work in a somewhat distracting context. Furthermore, our study confirmed that challenging demands may be experienced as hindering demands (and vice versa) depending on the context [63], as employees in the two groups perceived workload differently. The other reason may be related to the characteristics of the sample in the change group, as 76 percent of them were in professional, scientific, and technological occupations, whereas 75 percent of participants in the no-change group were healthcare and social workers. The characteristics of workload are quite different in these two groups; that is, a higher workload in the change group may entail more teaching or writing, whereas a higher workload in the no-change group may mean more patients or clients. Therefore, employees in the two studied groups may have perceived their workloads differently. In addition, the reason why workload was not significantly related to burnout for the change group but significantly associated with burnout in the no-change group could mean that for employees in the change group, WFH may have some advantages such as providing a more comfortable environment during the lockdown, more freedom of action, and reducing the time and costs of commuting. Therefore, the negative relationship between workload and burnout may be buffered by the advantages of WFH for employees in the change group. Second, in our study we discovered different relationships between the three sources of social support and wellbeing for the two groups of employees. Generally, our study proved that social support had different relationships with burnout and engagement for the change and no-change group. Specifically, for employees in the change group, leader support seemed to be more important for reducing burnout, as it was negatively related to burnout. Coworker support seemed to be more beneficial for promoting work engagement, as it was positively related to engagement. Furthermore, FWF was significantly related to both burnout and work engagement in the change group. Therefore, for employees who transitioned to more WFH, family support seemed to be a vital element to prevent burnout and promote work engagement. For employees in the no-change group, coworker support and FWF were both positively related to work engagement but showed no significant relationship with burnout. Additionally, leader support had no significant relationship with either burnout or engagement in the no change group. Overall, these results, to some degree, indicate that social support had a stronger relationship with work engagement, but neither were related to burnout in the no-change group. However, in the change group, social support seemed to be related to both work engagement and burnout. --- Limitations Our study contributes to the literature on working from home during the pandemic by investigating the relationships among demands, resources, and employee wellbeing in the change and no-change group. However, there are some limitations that should be noted. First, we utilized snowball sampling to collect our data, which is a nonprobability sampling technique where subjects are recruited through acquaintances. This approach provided convenience for our data collection but also led to nonrandom sampling, which might undermine the representativeness and generalization of the results. In addition, the occupations of employees in the change and no-change group in this study were not equally distributed. The distribution bias of our sample might influence the generalization of our results. To some degree, our sample reflected the reality of occupations that had a higher probability of working from home or remaining in the workplaces during the COVID-19 pandemic. However, the sample did not cover employees with a lower socioeconomic status, such as blue-collar workers, who had to work on the frontline during the pandemic. There have also been large and marked differences for COVID-19 health personnel in Norway compared with other countries. Among other factors, the pressure on health personnel may have been reduced due to the infection being controlled effectively: in 2020, Norway had one of the lowest death rates, with a total of 228 deaths among over 5 million inhabitants [64]. Therefore, even though the sample in this study was not equally distributed, we believe that the results may still have some implications for the future of work. Second, our study used a cross-sectional research design, which was insufficient for drawing conclusions about causal relationships between variables. Future research may consider using a longitudinal design to test the causal relationships among job demands, job resources, burnout, and work engagement within the context of the COVID-19 pandemic. Third, our study was based on self-report measures, which may lead to common method bias. Even though we tested for common method bias, and it was not an issue in this study, multiple methods of assessment should be considered in future studies. For example, future studies could consider utilizing objective measurements to evaluate employees' workload. --- Implications Our study indicated that there was a difference in the experience of workload and coworker support between the change and no-change group. Moreover, even though we found that all of the investigated job demands and resources were significantly related to employee wellbeing to some extent, there were some differences between the two groups regarding the importance of different job demands and job resources for employee wellbeing. Thus, practitioners working to increase wellbeing for both employees in the change and no-change group should focus their interventions on facilitating a healthy work-home balance, as WHC was the most important factor related to employee wellbeing in both groups of this study. This may have additional importance for employees WFH, as support from family was found to be the most important resource related to wellbeing. In addition, to reduce strain, interventions should focus on leader support, whereas to facilitate motivation, coworker support is important for the wellbeing of employees working from home. Thus, facilitating access to multiple sources of support and family-friendly policies seems to be an important focus area of healthy workplace interventions for employees in both working conditions. When it comes to the association between workload and wellbeing, the results from this study may be of particular interest to both researchers and practitioners. For employees still at work, workload seems to contribute to strain. However, even though employees WFH reported a higher workload, it seems to have motivational effects for them. On the one hand, this is an important finding for practitioners and managers occupied with the wellbeing of employees working in different work situations. On the other hand, this contributes to research on the JD-R model by demonstrating that the same job demand can be experienced as hindering or challenging for employees depending on their work situation. Moreover, it seems that for the different groups, different job resources are of importance for different outcomes. Therefore, practitioners should be aware of this when seeking to affect either motivational or health-impairing processes related to employee wellbeing. Researchers should also be aware of these mechanisms, especially when generalizing findings across work situations. --- Conclusions The COVID-19 pandemic swept across the world at the beginning of 2020 and fueled the fire of remote work. The aim of this study was to contribute to the literature on working from home during the COVID-19 pandemic. Employees who transitioned to working more from home and those who still went to work

After the outbreak of the COVID-19 pandemic, many employees were suddenly required to work more from home. Previous literature on working from home may not be applicable to this mandatory and overall change. In this study, we drew on the Job Demands-Resources (JD-R) model to explore the relationships between job demands (workload and work-home conflict) as well as resources (support from leaders, coworkers, and the family) and wellbeing (burnout and work engagement) in employees who still went to the workplace (no-change group) and employees who transitioned into working from home (change group) during the COVID-19 pandemic. Data were analyzed with multivariate structural equation modeling. The results indicate that work-home conflict was detrimental for employee wellbeing in both groups. Interestingly, the workload seems to contribute to work engagement for employees who worked from home. Regarding the resources, the three different sources of social support, leaders, coworkers, and family, were all related to employee wellbeing, but in different ways. It seemed that family support was most important for employees' wellbeing in the change group. This study presents implications for the wellbeing of employees in both the change and no-change group during the COVID-19 pandemic, emphasizing the importance of family-friendly policies.

may be of particular interest to both researchers and practitioners. For employees still at work, workload seems to contribute to strain. However, even though employees WFH reported a higher workload, it seems to have motivational effects for them. On the one hand, this is an important finding for practitioners and managers occupied with the wellbeing of employees working in different work situations. On the other hand, this contributes to research on the JD-R model by demonstrating that the same job demand can be experienced as hindering or challenging for employees depending on their work situation. Moreover, it seems that for the different groups, different job resources are of importance for different outcomes. Therefore, practitioners should be aware of this when seeking to affect either motivational or health-impairing processes related to employee wellbeing. Researchers should also be aware of these mechanisms, especially when generalizing findings across work situations. --- Conclusions The COVID-19 pandemic swept across the world at the beginning of 2020 and fueled the fire of remote work. The aim of this study was to contribute to the literature on working from home during the COVID-19 pandemic. Employees who transitioned to working more from home and those who still went to work were studied through the framework of the job demands-resources model and the conservation of resources theory in Norway, a country praised for its handling of the pandemic. The results of this study indicate that work-home conflict was an important job demand that was related to employee wellbeing in both groups. Workload showed different associations with wellbeing in the two groups. In addition, the three different sources of social support, namely, leaders, coworkers, and family, presented a positive relationship with employee wellbeing in different ways. Moreover, family support appeared to be more strongly related to wellbeing for employees in the change group. Therefore, employers may have a lot to gain from facilitating for family-friendly policies and social support from different sources in order to both reduce employee health impairment and increase their motivation. --- Data Availability Statement: The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy. --- Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. --- Conflicts of Interest: The authors declare no conflict of interest.

After the outbreak of the COVID-19 pandemic, many employees were suddenly required to work more from home. Previous literature on working from home may not be applicable to this mandatory and overall change. In this study, we drew on the Job Demands-Resources (JD-R) model to explore the relationships between job demands (workload and work-home conflict) as well as resources (support from leaders, coworkers, and the family) and wellbeing (burnout and work engagement) in employees who still went to the workplace (no-change group) and employees who transitioned into working from home (change group) during the COVID-19 pandemic. Data were analyzed with multivariate structural equation modeling. The results indicate that work-home conflict was detrimental for employee wellbeing in both groups. Interestingly, the workload seems to contribute to work engagement for employees who worked from home. Regarding the resources, the three different sources of social support, leaders, coworkers, and family, were all related to employee wellbeing, but in different ways. It seemed that family support was most important for employees' wellbeing in the change group. This study presents implications for the wellbeing of employees in both the change and no-change group during the COVID-19 pandemic, emphasizing the importance of family-friendly policies.

Introduction Settings throughout the United States and Canada continue to experience unprecedented levels of opioid overdoses driven largely by the increasing presence of synthetic fentanyl and related substances. This is true of Vancouver, British Columbia (BC), Canada, where in 2016 a public health emergency was declared in response to a rapid increase in overdose deaths; an 87% increase from 529 deaths in 2015 to 991 in 2016 [2]. Despite the existence of a range of harm reduction and treatment programs, estimated to have reduced total overdose fatalities from 1547 in 2018 to 983 in 2019, overdose rates remain high. The continued high overdose rate may be attributed in part to the stigma surrounding substance use, which creates a major barrier to accessing overdose prevention services and treatment programs [1,22]. The crisis continues to worsen during the COVID-19 pandemic as 1068 overdose fatalities were recorded in the first 8 months of 2020 alone [2]. Comorbid mental disorders are prevalent among people who use drugs (PWUD) and are known to be associated with suboptimal health and social service access, poor treatment retention and outcomes, as well as elevated overdose risk [3,9,10,20]. Among Canadians with a mental health disorder, those with a substance use disorder were over three times more likely than those without a substance use disorder to experience unmet mental health care needs [23]. Of particular concern is the high rate of post-traumatic stress disorder (PTSD) among PWUD. While the lifetime prevalence of PTSD in the general Canadian population is approximately 9.2% [24], the lifetime prevalence of PTSD among those with a substance use disorder is estimated to be between 26 and 52% [20]. PTSD is characterized by intrusions, avoidance behaviour, changes in mood and hypervigilance [8], all of which can impair goal-oriented thinking and behaviour, like pursuing supportive services and staying connected to services [17]. Additionally, recent data has indicated that a PTSD diagnosis almost doubles the overdose risk among people who use opioids [17]. Furthermore, many providers are untrained in trauma informed care (TIC), especially caring for PWUD [3], and can administer inappropriate care and perpetuate substance use related stigma, deterring PWUD from approaching services. While the ongoing overdose crisis has prompted efforts to improve access to health and social services and promote addiction treatment engagement [14], multiple factors can create service access barriers for PWUD with a PTSD diagnosis, heightening risks of poor treatment retention, outcomes and overdose death. While access to services for PWUD with varying secondary mental health diagnoses like major depressive disorder (MDD) [3], anxiety disorder [15], and other psychiatric conditions [6] have been well investigated, the impact of PTSD on PWUD's service access has not been well characterized. Therefore, we sought to explore the relationship between PTSD and difficulties accessing health and social services among PWUD in Vancouver. --- Methods Data were derived and combined from two ongoing and harmonized prospective cohort studies of PWUD in Vancouver, Canada: the Vancouver Injection Drug Users Study (VIDUS) and the AIDS Care Cohort to evaluate Exposure to Survival Services (ACCESS). Both cohorts have been recruited via snowball sampling, self-referral and street outreach since 1996 and have been described in detail elsewhere [21,27]. In short, VIDUS eligible participants are HIV-negative individuals who have injected illicit drugs at least once in the month prior to enrolment. ACCESS eligible participants are HIV-positive individuals who used illicit drugs other than or in addition to cannabis in the month prior to enrolment. At baseline and semiannually, both VIDUS and ACCESS participants complete a harmonized interviewer-administered questionnaire which gathers quantitative information on demographic characteristics, substance use and risk behaviours, drug use treatment and use of health services information. Participants from each cohort also provide a venous blood sample for serologic testing. All participants provide informed consent and are given a $40 (CAD) stipend after each study visit. The use of VIDUS and ACCESS data in the current study have been approved by the University of British Columbia's Research Ethics Board. Our analyses are restricted to VIDUS and ACCESS participants who completed at least two study interviews at two distinct time points between April 2017 and November 2018. The primary outcome of interest was the inability to access services, assessed by the question: "in the last 6 months, was there a time you were in need of a service (e.g., housing, counselling, police) but could not obtain it?". This definition has been used in previous studies [19,25] to assess service access barriers among PWUD and was coded dichotomously (yes vs. no). To assess for PTSD, each participant completed the PTSD Checklist for DSM-5 (PCL-5). This 20-item questionnaire evaluates PTSD symptoms experienced in the previous month. Each item is answered on a five-point scale (zero = not at all to four = extremely), and scores range from 0 to 80. Consistent with previous studies, a provisional PTSD diagnosis was defined as a PCL-5 score of 31 or higher [26]. After reviewing the literature [3,17,19], other characteristics considered as potential confounders included age (<unk>median vs. <unk>median), sex (male vs. non-male), ethnicity (white vs. Black Indigenous or other People of Colour [BIPOC]), and within the last 6 months prior to data collection interviews: homelessness (yes vs. no), incarceration (yes vs. no), sex work (yes vs. no), overdose (yes vs. no), employment (yes vs. no), alcohol binge (yes vs. no), daily use of prescription opioids (yes vs. no), heroin (yes vs. no), crack (yes vs. no), cocaine (yes vs. no), and methamphetamine use (yes vs. no). Our study aims to assess whether trauma, operationalized by a provisional PTSD diagnosis, is associated with the inability to access services for PWUD. Since repeated measures were available for each participant, we applied generalized estimating equation (GEE) logistic regression with logit-link function for the analysis of correlated data. By using an exchangeable correlation structure that adjusts for multiple observations for each participant, this method identified factors associated with the outcome across the entire study period. As a first step, a bivariable GEE analysis was used to determine factors associated with inability to access services in the last 6 months. In this analysis, the explanatory variables significantly associated with the outcome were PTSD, homelessness, recent incarceration, overdose, daily crack use and daily methamphetamine use. To fit the multivariable model, we began with all explanatory variables found to be associated with the outcome at p <unk> 0.10 in bivariable analyses, then generated a series of reduced models by removing each secondary explanatory variable one at a time. For each of these models we assessed the relative change in the coefficient for PTSD. The secondary explanatory variable that resulted in the smallest absolute relative change in the coefficient for PTSD was then removed. Secondary variables continued to be removed through this process until the smallest relative change exceeded 5%. Remaining variables were considered confounders and were included in the final multivariable model. Data analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC, USA) and all p-values were two-sided. --- Results In total, 810 individuals were included in this analysis, including: 488 (60.3%) males, 289 (35.7%) females, 10 (1.2%) transwomen, and 2 (0.4%) two-spirited individuals. Also included were: 343 (42.3%) white individuals, 338 (41.7%) individuals of Indigenous ancestry, 111 (13.7%) who identified as being a person of colour other than black, and 13 (1.6%) black individuals. Further, 645 (79.6%) of participants identified themselves as being "straight", 45 (5.6%) identified as being gay, 4 (0.5%) identified as being lesbian, 3 (0.4%) identified as being bisexual, and 3 (0.4%) identified as being twospirited, and 9 (1.1%) identified as belonging to another unspecified sexual minority group. The median age at baseline was 50.5 years [interquartile range = 41.8-56.1]. Overall, 117 (14.4%) of participants reported needing a service but were unable to access it during the study period, and 316 (39.0%) participants qualified for a provisional PTSD diagnosis. In bivariable GEE analyses (as shown in Table 1), factors positively associated with the inability to access services were PTSD, homelessness, recent incarceration, overdose, daily crack use and daily methamphetamine use, while older age was negatively associated with an inability to access services (all p <unk> 0.05). In a multivariable GEE analysis (as shown in Table 1), PTSD (adjusted odds ratio [AOR] = 1.69, 95% confidence interval [CI]: 1.12-2.55; p = 0.007) remained independently associated with an inability to access service, as did homelessness (AOR = 13.84, 95% CI: 8.96-21.36; p <unk> 0.001). --- Discussion Our study found that a significant proportion of participants self-reported difficulties accessing services, and over a third met the provisional PTSD diagnosis. In our multivariable analysis, inability to access services among PWUD was positively associated with a provisional PTSD diagnosis. One explanation for such findings is that PTSD can impair cognition, negatively affecting the ability to obtain services [17]. Therefore, impaired goal-oriented thinking coupled with the need for relief from emotional pain [7,17] can prompt PWUD to seek immediate self-soothing resources (e.g., use of opioids), rather than pursuing supportive services and staying connected in services, which often require long wait lists and specific criteria [4]. Similar to the current study, Beaulieu et al. [3] found that amongst VIDUS and ACCESS participants, major depressive disorder (MDD) diagnosis among PWUD was positively and independently associated with reported barriers to accessing health care. A possible explanation for this relationship proposed by the authors was service provider displays of stigma. As the current study involved the same cohorts as that of Beaulieu et al. [3], the same explanation may extend to the current study. Displays of stigma, related to mental illness, substance use and homelessness, may deter clients from reaching out for services or staying in touch with services, despite needing such supports. As well, as suggested by Beaulieu et al. [3], service providers may be unfamiliar with how to assist PWUD who are diagnosed with PTSD, and their lack of confidence in serving this population results in inconsistent support and compassion fatigue. Further, avoidance of addressing difficult issues, like trauma, may be interpreted as mistreatment or stigmatized views by clients, which may further deter PWUD from reaching out or staying connected with much needed supports and services. Thus, critical next steps involve further training and education for service providers on the issues of drug use and trauma related mental illness, allowing them to better respond to and work with clients from backgrounds of trauma [11]. A study by Morrissey et al. [18] found that 31% of female PWUD with severe PTSD symptoms given TIC for 12 months had their trauma symptoms reduced to moderate or less, compared to only 16% given non-TIC. Furthermore, clients who receive TIC report feeling safer and better able to self-advocate in service settings [13], making clients more likely to access and remain connected to services provided in a TIC context. Such changes have potential to create a more suitable emotional climate to better serve PWUD facing these challenges, with the broader aim of reducing perceived barriers to accessing the support that they need. Additionally, past reviews suggest that PWUD commonly report an inability to access community mental health services like counselling [7], consistent with the current study. This could be due to the lack of funding for and scarcity of such services, or could be attributed to the limited amount of services offered by providers, which was highlighted as a primary barrier to service access in a study involving PWUD in Edmonton, Canada [12]. Those seeking services may have been able to gain initial access; however, they were not granted the appropriate amount of services to meet their needs, whether it be frequency or duration of services, therefore creating This indicates a need to expand community mental health services provided to accommodate clients with higher service needs, as PWUD living with PTSD generally need mental health care more frequently and longer term than those in the general population [15]. The current study is limited by the fact that participants in our study were not randomly selected, which limits the generalizability of our results to those close in characteristic to the study sample. Further, we relied on self-reported data for both measures of the primary outcome and predictor variable, and as such, these data are subject to response biases like social desirability and recall biases [5,16]. Our data was also observational, therefore no causal relationship between variables can be drawn. Finally, though we did adjust for potential confounders in our analyses, additional unmeasured factors may have affected the results. --- Conclusion Overall, we found that PWUD are more likely to experience the inability to access services if they are experiencing PTSD. Our study suggests that service providers may benefit from education and training in TIC approaches to enhance access for PWUD with PTSD, as this client population might have difficulties navigating the system due to the mental and emotional difficulties that accompany PTSD. At a broader level, efforts should be made to expand services to account for more complex client needs and should accommodate those with concurrent mental health disorders who have higher service needs. --- Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. --- Abbreviations PTSD: Post-traumatic stress disorder; PWUD: People who use drugs; TIC: Trauma informed care; PCL-5: PTSD Checklist for the DSM-V; VIDUS: Vancouver Injection Drug Users Study; ACCESS: AIDS Care Cohort to evaluate Exposure to Survival Services; MDD: Major depressive disorder; GEE: Generalized estimating equation Authors' contributions TK, AG, and WL conceived and designed the study and drafted the manuscript. HD performed the statistical analysis and provided substantial revisions to the manuscript, as did KH, WL and MJM. All authors have read and approved the finalized version of the manuscript. --- Declarations Ethics approval and consent to participate All participants provided informed consent and VIDUS and ACCESS cohorts have been approved by the University of British Columbia's Research Ethics Board. --- Consent for publication Not applicable. --- Competing interests The authors declare that they have no competing interests. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: Settings throughout the United States and Canada are contending with high rates of drug-related overdose. This in turn has prompted efforts to more effectively engage people who use drugs (PWUD) in treatment and care. However, while co-morbid mental disorders are prevalent among PWUD and can undermine access to services, the impact of post-traumatic stress disorder (PTSD) on service access is not known. Therefore, we sought to assess whether PTSD is associated with difficulties accessing health and social services among PWUD in Vancouver, Canada. Methods: Survey data was derived from two prospective cohorts of PWUD in Vancouver, Canada for the period of April 2017 to November 2018. PTSD was assessed using the PTSD Checklist for the DSM-V (PCL-5). Generalized estimating equations (GEE) was used to estimate the relationship between PTSD and self-reported inability to access health and social services, after adjustment for confounders. Results: Among 810 participants included in our analysis, 316 (39.0%) participants qualified for a provisional PSTD diagnosis, and 117 (14.4%) reported difficulties accessing services. In a multivariable GEE analysis, a PTSD diagnosis (adjusted odds ratio = 1.69, 95% confidence interval: 1.12-2.55) was independently associated with difficulties accessing services. Conclusions: We found high rates of PTSD and self-reported difficulties accessing services among PWUD in Vancouver, as well as a positive association between PTSD and difficulties with service access. These findings highlight the need for trauma-informed approaches to service delivery for PWUD, as well as enhanced provider training specific to PTSD.

Passel, 2009). Despite this growth, services in health care and education have not expanded in parallel to meet the needs of this burgeoning, diverse population. This limited response has likely contributed to the disparities now found in health care, mental health, and educational opportunities available and accessible to Latinos residing in the United States. Disparities are defined by The National Healthcare Disparities Report (2010) as health conditions that are unequal to some degree, including factors that are associated with differential rates of disease for a subgroup or population (see Carter-Pokras & Baquet, 2002 for analysis of existing state and federal definitions of disparities). Generally, ethnic minority populations in the United States experience health disparities, including a reduction in opportunities to access and benefit from quality health care services. For example, fewer than 1 in 20 Latino immigrants with mental disorders use services of mental health specialists ( U.S. DHHS, 2001) and advances in medicine are less likely to reach Latinos ( Reyes, Putte, Falcon & Levy, 2004). Additionally, disparities have been documented in parental access to services regarding their child's mental health ( Alegr<unk>a et al., 2004) and in representation in center-based childcare and prekindergarten programs that promote school readiness ( Karoly & Gonzalez, 2011;Mather & Foxen, 2010). To overcome disparities, researchers have suggested that cultural adaptations of existing treatments would help address the underutilization of services by ethnic minority populations ( Smith, Domenech Rodriguez & Bernal, 2011). Several models have been proposed to help service providers adapt existing programs or develop new ones that are tailored to meet the specific needs of diverse families. Lau (2006) presented a theoretical framework for determining when and how to develop a cultural adaptation of a psychological treatment. The model contains two key elements; specifically, engagement, which refers to how well an intervention can successfully reach and engage potential participants, and outcomes, which refers to how well the intervention meets target goals. Lau (2006) proposed that participants' engagement in an intervention rests on its social validity, or the degree to which recipients believe a service is useful. Barrera and Castro's (2006) expanded model of cultural adaptation shows that engagement can consist of a number of components, including awareness of treatment availability, entry into treatment, participation in treatment activities, and completion of treatment. Social validity, defined as the "perceived acceptability and utility, perceptions that might be influenced by cultural worldviews and the practical realities of life circumstances," ( Barrera & Castro, 2006, p. 312) can influence engagement at all levels. --- Gaps in the Literature Although some research has demonstrated the effectiveness of culturally adapted treatments ( Domenech Rodr<unk>guez, Bauman, & Schwartz, 2011;McCabe, Yeh, Garland, Lau, & Chavez, 2005;Reese & Vera, 2007), existing studies generally do not explain the process of how cultural factors influence treatment engagement by the participants. Based on reviews of existing treatment outcome research, it appears that when a treatment is closely aligned to the cultural worldview of the client, the more likely it is to be successful ( Smith et al., 2011). Scholars examining culturally adapted treatments disagree as to the degree to which adaptations should be implemented from the outset (e.g., Bernal, Bonilla & Bellido, 1995) or if evidence-based practice should be modified only when existing treatments have been shown to be ineffective. There seems to be greater agreement that intervention delivery should at minimum occur in the client's preferred language ( Smith et al., 2011). Presently, we are lacking studies that describe and critique the process of adapting a treatment to be relevant for Latino populations. Research is needed to examine how specific cultural factors, such as acculturation, interact with other contextual barriers faced by Latino families, such as economic factors, to influence treatment engagement. Producing culturally adapted treatments for a range of issues specific to Latino populations that are available in Spanish is also a pressing need ( Calzada, Fernandez & Cortes, 2010). Although multiple meta-analyses suggesting that evidence based practice (EBP) using cultural adaptations, specific to the client's background and preferred language, produce stronger effects than programs without such adaptations, the availability of these types of programs is insufficient ( Griner & Smith, 2006;Smith et al., 2011). The present study seeks to address these gaps in several unique ways. First, we describe a process used to develop a culturally adapted treatment for promoting parent involvement by Latino parents in Head Start preschool programs. Second, we pilot and implement the new intervention while simultaneously examining engagement factors affecting participation in the treatment. To this end, we measure parent perceptions of contextual barriers that likely influence our participants' program attendance in addition to their level of acculturation before beginning the intervention program. By simultaneously considering how barriers to participation are related to the overall acculturation of the parents in our study, we seek to offer a more complete picture of how acculturation may interact with other engagement factors. We examine the role of acculturation in treatment engagement with a sample of relatively recent-arriving Latino families, seeking to determine whether a culturally adapted treatment is useful for this population. This goal was selected because research has shown that families with low levels of acculturation to the U.S. are least likely to access quality health services ( Lara, Gamboa, Kahramanian, Morales & Bautista, 2005). In the next section, we describe the conceptual model used to guide this research study. --- Conceptual Model To inform our research, we developed a theoretical model that is similar to those proposed earlier by Barrera and Castro (2006) and Lau (2006). We selected these models because of the explicit consideration by these authors of factors that would promote engagement, as well as the emphasis on planned accommodations that would be refined through pilot work. Figure 1 depicts multiple factors that could influence engagement in an intervention designed to promote parent involvement among low-income Latino parents of preschool children attending Head Start. Specific factors in our model include barriers to participation, perceived benefits of the program, perceived economic stress, and parents' level of acculturation. Considering acculturation as an engagement factor is a unique and important feature of our model because although participants in our intervention were recent immigrants to the U.S., they likely have different opportunities and resources available to them to learn about U.S. culture. Additionally, measuring acculturation directly, as opposed to utilizing proxy variables such as length of residency in the United States or English-speaking ability, is a more precise approach to assessing the importance of this construct to treatment engagement ( Lara et al., 2005). We sought to test the influence of each of these engagement factors on attendance in our program and whether parents' level of acculturation would influence their experience of barriers to treatment, or alternatively, their perception of benefits to treatment. Figure 1 also summarizes the planned adaptations that were developed through a series of steps, including a recommended brief pilot of the proposed new intervention ( Barrera & Castro, 2006). Before describing in greater detail the adaptation process, we briefly review extant literature involving treatment engagement for immigrant populations. --- Factors That Influence Engagement in Treatment The call for greater development and use of culturally adapted treatments is driven by the recognition that an individual's culture influences treatment engagement ( Barrera & Castro, 2006;Lau, 2006). One notable factor that has been measured among culturally diverse populations, particularly immigrant groups, is acculturation. Scholars view acculturation as involving both psychological changes in attitudes, values, and behaviors that result when an individual or group has contact with a new or different culture ( Berry, 1980( Berry,, 2003( Berry,, 2006;;Zea, Asner-Self, Burman, & Buki, 2003). Theoretical models of understanding acculturation for immigrants have moved toward a recognition of the importance of the bidimensional process that occurs as the individual navigates a new culture while also maintaining or modifying features of one's native culture ( Harwood & Feng, 2006;Lara et al., 2005). In Berry's (1980) bidimensional model, acculturation for immigrant populations is viewed best as a continuum on two dimensions-individuals range on their capacity to maintain their native cultural competence and also range in their adaptation to the culture of the new society. In reviewing literature involving immigrant and low-income samples, some studies reported how acculturation may influence parent engagement in complex ways. For example, parents hold beliefs regarding the importance of education and have respect for teachers; however, one study found they did not expect to be involved with the school directly ( Mapp, 2003). In fact, parents viewed parent involvement as possibly disrespectful to the teacher and school. Also, a lack of familiarity with expectations of U.S. schools could influence parents' efficacy about school involvement. Research with Latina mothers demonstrated that lower acculturation to the United States was associated with less knowledge about school activities and more barriers to parent involvement; however, parents with lower acculturation reported higher levels of efficacy and expectations for their children's educational attainment than Latina mothers with higher acculturation ( Moreno & Lopez, 1999). Finally, Mexican American mothers reported that acculturation differences between parents and children, separation from extended family, discrimination against immigrants, and concerns about legal status negatively impacted their parental involvement ( Leidy, Guerra & Toro, 2010). Researchers have also examined the influence of barriers to treatment on engagement among low-income, ethnic minority samples, apart from cultural factors such as acculturation ( Karoly & Gonzalez, 2011). For example, a study of low-income preschool parents found that contextual barriers, such as work schedules, time, and transportation availability significantly decreased intervention attendance by parents ( Mendez, Carpenter, LaForett, & Cohen, 2009). In another low-income, ethnically diverse sample, similar barriers emerged including work schedules, caring for children and elderly parents, and lack of transportation ( Mapp, 2003). Finally, parents report barriers such as negative relationships with the school, including perceiving their school involvement as unwanted ( Baker, 1997). Less attention has been given to a related factor, perceived benefits of treatment, which may influence treatment participation. In general, it is predicted that parents who expect to receive greater benefits from a program will be more likely to participate than parents who expect fewer benefits. In a community-based intervention study, parents' perceived benefits of a program were associated with project participation, whereas perceived barriers were negatively associated with project participation. However, this was only true for parents who initially intended to participate ( Spoth & Redmond, 2000). Greater study of the role of perceived benefits for participation among immigrant samples is warranted, as these factors could inform the social validity of an intervention. As noted by Lau (2006), when parents perceive a service is useful to them, the social validity is high and participation may increase. --- Cultural Adaptation of an Intervention for Head Start Latino Parents The increase of Latinos nationally is also reflected in the number of parents enrolling their children in Head Start. Latinos are now one of the largest groups served by Head Start programs (36%) and 28% of children attending Head Start live in households where English is not the primary language spoken ( Aikens et al., 2010). Therefore, there is a significant need to develop programs that are socially valid for Latino families, especially in the area of parent involvement. Head Start defines any child raised in a bilingual household as a Dual Language Learner student, and seeks to provide services in particular for these families. Based on an existing relationship with our local Head Start provider, we formed a partnership to develop and implement a culturally adapted parent involvement program. We worked closely with teachers, program administrators, and parent representatives to adapt an existing program for use in this Head Start setting and decided to target all parents who reported speaking a language other than English at home, the majority of whom were Latinos. This approach was consistent with Head Start policies, which articulate that parent services should be open to all interested families; however, our conversations primarily focused on cultural adaptations that would be appropriate generally for immigrant parents and specifically for Latino parents. We selected an intervention for consideration as a culturally adapted treatment that was available in the empirical literature and was previously used effectively in Head Start programs with African American families. The original program, called The Companion Curriculum (TCC; Mendez, 2010), promotes parent involvement at home and within the school setting using a series of lessons and materials to promote parent-child interactions. In considering possible adaptations that were needed, we identified the following issues primarily by having informal conversations with a Latina center director, bilingual parents, family service workers at Head Start, teachers, and by observing current parent meetings offered by the program. Following suggestions provided by Auerbach (2004), we discussed how small, intimate group meetings held in the school building would best allow Latino immigrant parents to feel comfortable sharing their true concerns. We also determined that information should be shared with parents using bilingual speakers with similar cultural backgrounds ( Auerbach, 2004). Several individuals from the leadership of our Head Start program suggested that combining parenting lessons with English as Second Language instruction into one program would increase the social validity for immigrant parents. Therefore, we incorporated materials typically found in English as Second Language (ESL) courses as part of the new intervention. Embedded in these English lessons was information about United States culture to help parents navigate new situations such as school, doctor's offices, or local transportation. By encouraging parents to have discussions in both English and their home language about a wide range of topics, it was our intention that the culturally adapted intervention would meet the needs of parents along a continuum of acculturative experiences, thereby minimizing this access barrier. For example, parents with more familiarity with U.S. schools could assist other parents; however, other parents may have more experience with the health care system. The second major cultural adaptation was to translate nine Parent Excellence lessons and a summary session into Spanish. Handouts for each lesson were provided in both the home language and English to reinforce the exposure to two languages. We involved multiple translators from several different Latin American backgrounds in order to select intervention content and language that might be meaningful and understood by parents from a variety of Spanish-speaking countries. A third adaptation involved hiring and training bilingual instructors to teach the Parent Excellence Lessons in either Spanish or English. The bilingual center director was a forceful proponent of providing parents opportunities to practice writing and communicating in English, to eventually encourage them to be full participants in their child's education. To provide scaffolding to learn new skills, all of the key concepts from each lesson were first reviewed in Spanish and then in English by the class leaders. Parents were encouraged to speak and promote literacy with their children in both their home language and English. To build relationships between care providers and participants, we had the bilingual facilitator make reminder calls to parents using Spanish before class meetings and also assist parents with any reported access barriers, such as transportation or child care needs. A minimum of one hour of training on each lesson, and review of phone calls and previous sessions, occurred weekly for all class facilitators throughout the duration of the program. Supervision was provided by the study investigators. In the majority of class meetings, we used two trained instructors with a ratio of about 8-10 parents for every one instructor. After completing our adaptation phase designed to enhance the social validity of the intervention, we used contacts from the family service worker to recruit a sample of parents to conduct a pilot of four TCC lessons and the ESL content. Following the 4-week pilot, we asked 17 parent participants to discuss strengths and weaknesses of the program and whether they or others like them would participate in a longer version of the intervention consisting of similar lessons. Feedback showed parents' unanimous interest in continuing to hold sessions at the school and to offer this program again. They reported that the ESL content and parenting information were helpful to their family but requested that Head Start provide a longer program on multiple days that would give them more time for conversations and learning. Parents also suggested that the program should be offered multiple times during the school year. Following these suggestions, we increased the intervention dosage from one hour per week to three hours per week and held meetings twice weekly. Lastly, parents who participated in the pilot reported that some lessons were easier to understand and use at home with respect to the content than other lessons. As outlined in Table 1, the original content of the intervention followed a particular order, with the opening session focusing on play as a learning tool for child development. The Latino parents offered to us that storytelling was a favorite concept from the curriculum, and therefore it seemed appropriate to move this lesson earlier in the intervention sequence. Other examples of cultural adaptations included featuring children's books that were available in Spanish and English with a range of characters in the Reading Together lesson or encouraging families to share stories, pictures, and/or practices of their home country with their children in the Building an I Can attitude lesson. In this way, parents were afforded the opportunity to incorporate the Latino culture into activities designed to promote school readiness and success in preschool for their children. Table 1 provides an overview of each specific session of the program, with the concepts designed to promote three components of school readiness for children (literacy, numeracy, and socioemotional competence). --- Table 1 is omitted from this formatted document. The outcome of these steps was a 12-week intervention consisting of two weekly 90-min classes, one using a Parent Excellence Lesson and one focusing on ESL instruction using a workbook of English vocabulary and grammar (for intervention materials, contact author). Trained bilingual facilitators taught each lesson in both English and Spanish as needed. Parents received doublesided lesson handouts in both English and Spanish, as well as colorful, child-friendly materials (e.g., puppets, alphabet letters) designed to illustrate the theme and to foster parent-child interaction at home. The ESL portion of the class incorporated information (e.g., reviewing vocabulary in English and helpful phrases) to assist parents with navigating services, such as how to use the pharmacy, or United States schools. Importantly, the sessions were all held at the child's school, a familiar location to the family members. --- Present Study This study examines how perceived barriers and potential benefits affect participation in a culturally adapted intervention for Latino parents. We hypothesized that Latino parents would report experiencing a range of barriers to treatment participation, as these factors have been previously identified in treatment studies involving low-income populations ( Kazdin, Holland, & Crowley, 1997;Mendez et al., 2009). We hypothesized that because of our cultural adaptations and changes after the pilot, the social validity would be adequate, meaning that parents would endorse a range of benefits associated with participation in our program, such as learning English or increasing their parent involvement. Because of the cultural adaptations made to the treatment, (e.g., use of bilingual facilitators and materials), we expected that parents ranging in level of acculturation would participate. Based on mixed findings in the literature ( Lara et al., 2005) and the lack of studies that use a bidimensional model of acculturation, we tentatively hypothesized parents with low levels of US cultural competence would perceive greater benefits associated with treatment than highly acculturated parents. Finally, we examined how parent-reported barriers would influence participation, above and beyond level of acculturation, perceived benefits of the program, economic stress, and length of residency in the United States. By first controlling for these variables, our investigation sought to determine how barriers that might be more modifiable would engagement by Latinos families in culturally adapted treatments. --- Method --- Participants Latino parents of children enrolled in Head Start programs who expressed interest in an adult literacy and parenting program were eligible for this study. Parents in our study lived in suburban areas on the outskirts of a major metropolitan area in the Northeast. A total of 54 Latino parents (91% female) participated in the study, representing a variety of countries of origin including Ecuador (30.91%), Mexico (21.82%), Brazil (20%), Honduras (14.55%), Peru (5.45%), Colombia (3.64%), and Puerto Rico (1.82%). The participants' mean reported monthly income was $1,052 ( n = 40, range, 0-2,400) with a median length of residence in the United States of 6 years (range 1-20). Ninety-four percent of the sample reported living in the United States for 10 or fewer years. Participants had a child 3.63 years of age on average, which is consistent with the target age group for Head Start children (ages 3-5). The majority of the parents (53.7%) were married or living with a significant other, 35.2% were single, and 11% were divorced or separated. The majority of parents (81.5%) did not attend school in the United States. Of the six parents who reported attending school in the United States, two earned a high school diploma or GED, two completed job training or vocational school, one completed 9th grade, and one did not report the level of schooling completed. --- Study Procedures This study was conducted in partnership with a community action agency offering Head Start services over a three-year period in accordance with approvals from our university IRB. The study procedures were also approved by the Head Start parent policy council and the agency director. We rotated offering the program at three centers during this time, once each year in the fall and the spring, generally following the Head Start calendar and avoiding weeks when school was not in session (e.g., winter break). Parents were recruited to participate using multiple recruitment strategies including speaking at parent orientation sessions during the beginning of the school year, attending parent meetings each month, sending home information sheets in both native and English language with children identified by the school as Dual Language Learners, asking the school staff and prior parent participants to refer parents to the program, and meeting with all school staff. Study facilitators presented parents with written information in both Spanish and English about the study and consent forms. A bilingual research assistant explained project goals and objectives to the participants and read consent forms orally. Upon enrollment in the study, parents completed self-report measures two to three weeks before participating in the adapted intervention and again at the conclusion of the program. All parent data were collected via individual interviews completed in person or via telephone with a bilingual interviewer who had no contact with parents enrolled in the classes. Parents received a $30 dollar gift card for participating in each assessment. Parents kept all the intervention materials and workbooks. Parents were not provided financial compensation to attend the intervention classes. --- Measures --- Perceived barriers and benefits The Barriers/Benefits measure ( Mendez et al., 2009;Mendez, 2010) was developed in prior intervention research studies to assess access barriers that influence families from low-income backgrounds. Barriers assessed include the following: No need for parenting program, lack of child care, not comfortable with talking about parenting with others, work schedule conflict, concerns that your family or friends would disapprove, health related problems, religious activities, transportation, night classes, too tired, or "other" barriers not captured by the previous categories. Parents indicate whether they perceived each item as a potential deterrent from their participation with a rating scale of 1-4 (1 = Definitely Yes, 2 = Probably Yes, 3 = Probably No, 4 = Definitely No). Items are reverse coded for scoring so that higher scores represent more perceived barriers. Benefits assessed include the following: Increasing your involvement with the Head Start program, improving your parenting skills, helping you communicate with your child's teacher, improving your child's understanding of letters and numbers, improving your child's behavior, providing you with educational activities to do with your child, improving your knowledge of child development, meeting new parents, reducing stress in your life, and understanding things about your child's schooling. Parents rated items on a scale of 1-4 (1 = Not at all beneficial, 2 = A little beneficial, 3 = Somewhat beneficial, 4 = Extremely beneficial). A higher total score signifies more perceived benefits. --- Acculturation The Abbreviated Multidimensional Acculturation Scale (AMAS-ZABB; Zea et al., 2003) is a 42item measure that assesses aspects of acculturation in both the native culture and U.S. culture. For this study, the Native Cultural Competence, U.S. Cultural Competence, Native Language, and U.S. Language subscales were used in analyses. Cultural competence includes knowledge about the culture and the ability to function in it. Items on this scale include familiarity with national heroes, TV shows, pop culture, and politicians. Language competence assesses how well the reporter speaks a given language in different situations, such as school, work, with friends, on the phone, and in general. Most of the validation work has focused on multiple samples of Spanish-speaking populations, and good evidence has been obtained in support of construct validity. In our sample, there was a high correlation between U.S. language and U.S. competence ( r =.80, p <unk>.01); therefore, we determined that use of a composite score in analyses was warranted. The composite score for U.S. acculturation was created by taking the average of the U.S. Language and U.S. Competence subscale scores, which showed good reliability with a Cronbach's alpha of.95. Cronbach's alphas for Native Language and Native Competence were.96 and.93, respectively. --- Demographics and economic stress Parent reported demographic information included duration of residency in the United States, country of origin, number of children, number of people living in home, marital status, occupation in the United States, education, income, and economic stress. Economic stress was measured using items developed by Conger and colleagues (1992) that assess the degree to which parents' perceive their income as sufficient to meet their needs. --- Attendance Attendance was measured by the total number of classes parents attended. Some intervention sites differed slightly in number of classes offered, primarily because of snow days and other calendar differences across the school year; however, all sites provided instruction using the same ESL workbook and nine Parent Excellence lessons. To make the attendance variable, a percent attendance variable was created by dividing the total number of classes the parent attended by the number of classes offered at that location. --- Data Analysis We computed means and standard deviations for all study variables. We used paired t tests to statistically compare the levels of acculturation or competence in both the native and U.S. cultures. Bivariate correlations were run to inform the use of hierarchical regression. We used this type of regression model to determine how much variance in program engagement was associated with reported barriers to treatment, after controlling for length of residence in the U.S., benefits, acculturation, and parent economic stress. Because U.S. language and U.S. competence were very strongly associated we averaged these scores to make a composite score, which was used for all remaining analyses. In step one, we entered length of residency in the United States. In step two, acculturation indices (native competence, native language, U.S. language/competence composite) and parent perceptions (perceived economic stress and perceived benefits of the program) were entered into the model. Lastly, the barriers-to-treatment variable was entered into the regression. --- Results Table 2 reports the average score for each barrier and benefit item on the survey in descending order, starting with the most frequently endorsed barrier. The overall mean number of perceived barriers in this sample of Latino immigrant parents was 1.81 with a range of 1.00-3.08 (see Table 2). The barrier endorsed most often by the sample was "Difficulty finding childcare." The least often barrier endorsed was "Not comfortable talking about parenting with others." The overall mean number of perceived benefits reported by the sample was 3.54 with a range of 3.2-3.83 (see Table 2). The benefit endorsed most often by the sample was "Understanding things about your child's schooling." The least often benefit endorsed was "Helping you communicate with your child's teacher." Table 2 is omitted from this formatted document. Next, means and relations among the four subscales from the AMAS-ZABB measure of acculturation were examined (see Table 3). Paired samples t tests showed significant differences between all acculturation scales, indicating that the sample reported the highest scores for measures of Native Language, followed by Native Competence, with U.S. Language and U.S. Competence yielding the lowest subscale scores. Table 3 is omitted from this formatted document. We examined the bivariate correlations among the subscales from the AMAS-ZABB and length of stay in the United States. Native Competence and U.S. Competence were positively correlated ( r =.29, p <unk>.05), but to a much lesser degree than U.S. Language and U.S. Competence ( r =.80, p <unk>.01). Length of residency in the U.S. was positively correlated with U.S. Language ( r =.29, p <unk>.05) and negatively correlated with Native Competence (r = -.36, p <unk>.01). No other correlations were significant. Next, the associations between acculturation constructs and the barriers/benefits variables were examined. Results showed significant correlations between Native Competence and barriers ( r = -.29; p <unk>.05) and between Native Competence and perceived benefits ( r =.30; p <unk>.05). Correlations involving length of residence or the measures of U.S. acculturation with perceived barriers and/or with perceived benefits were not significant. Following this step, we compared program attendance and acculturation constructs and found no significant relations. However, we did find that barriers were negatively related to attendance ( r = -.27, p <unk>.05). Table 4 reports the results from the hierarchical regression model. After controlling for acculturation variables and parent perceptions, there was a significant relationship between expected barriers-to-treatment and attendance at the program (<unk> = -0.32, p <unk>.05). Table 4 is omitted from this formatted document. --- Discussion The purpose of this study was to examine how implementing a culturally adapted treatment might reduce barriers to participation in treatment by immigrant Latino parents of preschool children. Overall, the results showed that specific barriers to treatment do interfere with participation by Latinos, above and beyond the influence of acculturation indices, perceived economic stress, and anticipated benefits of the program. These findings underscore the importance of considering multiple factors, including barriers and acculturation differences, when developing and implementing culturally adapted treatment programs. From a public health perspective, there have been calls for "outreach programs" targeting immigrant populations with low levels of acculturation, as some studies have documented limited access to health care for this group ( Lara et al., 2005, p. 385). Our results show that by anticipating engagement factors for a recently arriving population of Latino immigrants, community-based programs can be successful in meeting the diverse needs of this population. In this study, we planned a series of accommodations to account for and deliberately recruit parents of with low levels of U.S. Competence and Language including (a) providing ESL classes, (b) offering the materials in both Spanish and English, and (c) training and providing bilingual facilitators to deliver the intervention. Using supports in Spanish to allow parents to develop English language skills and learn about educational approaches used in the United States, our program deliberately encouraged participants to maintain their native cultural competencies (e.g., discussing how to teach children to read using Spanish or English books; storytelling about Latino culture using Spanish) while expanding their U.S. cultural competence. Our study found that parents reporting high levels of Native cultural competence endorsed more perceived benefits of the intervention. We interpret these results as evidence that the scaffolding experiences featured in our culturally adapted program were instrumental in facilitating participation for those parents who were knowledgeable about their own culture but perhaps had less familiarity with the United States, perhaps because of their limited time living in this country. Our conceptual model emphasizing the importance of considering acculturation with Latinos was empirically supported. Because we found that higher levels of native cultural competence were associated with fewer perceived barriers to treatment, perhaps these individuals are more inclined to participate in a program that could help them adjust to life in the United States and learn information about navigating this new culture. We might consider these individuals as adopting the integrative approach to biculturalism, where individuals retain their native cultural competence while seeking to add new skills and values from the receiving culture ( Berry, 2006). Alternatively, parents who possess other avenues for learning English and who already perceive themselves as aware of U.S. cultural practices may see less need for this type of service, as compared with parents who have limited resources to assist with overcoming language and cultural barriers. Another distinctive feature of this study involves the measurement of acculturation using the recommended bidimensional approach ( Lara et al., 2005), such that we could examine both native cultural competence and emerging U.S. competencies for this sample of fairly recent immigrants. In our study, length of residence related positively to U.S. language competence and negatively to Native cultural competence, but was not related to other aspects of acculturation. However, parents with higher levels of Native cultural competence perceived greater benefits associated with the program. Perhaps these parents felt a strong appreciation for a parenting program for Spanish-speakers and felt a connection to presenters who shared their language and cultural background as noted by Auerbach (2004). Future studies can examine this relation more precisely by investigating potential mediators among acculturation and benefits and usage of interventions, such as parental self-efficacy or ethnic identity/cultural pride. Similar to extant literature on engagement of Latino parents within school settings ( Auerbach, 2004( Auerbach,, 2009;;Shapiro, DuPaul, Barnabas, Benson, & Slay, 2010), we endorse the importance of establishing a collaborative academic-community partnership before conducting research and intervention studies. Our partnership involved a community of recent immigrants, members of our local Head Start program, and the university members of the research team, including Latino students and faculty. The partnership was collaborative as the goals were determined with input along the way from community members ( Suarez-Balcazar et al., 2004). Gathering input from these stakeholders during the adaptation and pilot was an important strategy for ensuring the social validity of this culturally adapted treatment. Interestingly, because of the partnership, families outside of the Latino culture who were also recent immigrants felt welcome attending some of the intervention sessions alongside of these Latino families. Therefore, it may be that this intervention is sustainable as it falls within the mission of Head Start family engagement services being open and inclusive of all interested immigrant families. To advance the health and wellbeing of Latino children, barriers to treatment must continue to be studied and addressed to maximize access to services for underserved populations. As noted by Karoly and Gonzalez (2011), children of immigrant parents remain less likely to enroll in early childhood programs as a result of access and affordability issues, which further restricts opportunities for their parents to participate in center-based interventions. Reviews also suggest that greater barriers to accessing health care are found among immigrant families with less acculturation (e.g., Lara et al., 2005). We recommend that future research test more complex theories regarding how factors associated with acculturation for immigrant and Latino populations influence their attitudes and/or decisions regarding how to access services. For example, it may be that parents' work environment may contribute positively to their emerging acculturation to U.S. culture and increasing English language ability. Also, parents who are employed outside the home and use child care benefit from interactions with staff at child care centers by understanding more about child development and U.S. school systems. However, working parents likely experience barriers associated with their work schedules that may impact their willingness or ability to access health care services or educational programs. These issues represent areas of inquiry for future study. --- Limitations Despite the strengths of our study, we must acknowledge several limitations. We measured barriers to treatment just before treatment implementation; therefore, this measure reflected the barriers that parents expected to face in participating in the intervention rather than the actual barriers during the 12 weeks. It may be that

Literature to date has not yet included a comprehensive examination of barriers to participation in treatment for Latino populations, incorporating the role of acculturation. This study reports on the process of developing and implementing a culturally adapted treatment for promoting parent involvement by Latino parents in Head Start preschool programs and examines engagement factors affecting participation in the treatment. Results show that Latino parents with higher native cultural competence perceived more benefits to the treatment and reported fewer barriers than parents with lower native cultural competence. Also, the total number of barriers to participation that parents perceived before treatment was negatively associated with treatment participation, above and beyond the influence of acculturation factors, parent perceptions of economic stress, and perceived program benefits. Results are discussed in terms of how to establish effective community-based mental health and educational outreach programs to promote engagement with Latino families. The importance of using bidimensional models of acculturation within research designs with Latino populations and ideas for future research are discussed.Latinos are presently the largest ethnic minority group in the United States. According to the 2010 Census, more than 50 million Latinos live in the United States, accounting for 16.3% of the total U.S. population and 23.1% of children under the age of 18 ( U.S. Census Bureau, 2010). The Latino population represents one of the fastest growing ethnic groups in the United States, showing a 30.3% increase from 2000 to 2010, compared with a 10.6% increase for African Americans and a 1.1% increase for non-Hispanic Whites ( Pew Hispanic Center, 2008; U.S. Census Bureau, 2010). As of 2007, more than half of Latino children were classified as "second generation," meaning that they were born in the United States to immigrant parents ( Fry &

in center-based interventions. Reviews also suggest that greater barriers to accessing health care are found among immigrant families with less acculturation (e.g., Lara et al., 2005). We recommend that future research test more complex theories regarding how factors associated with acculturation for immigrant and Latino populations influence their attitudes and/or decisions regarding how to access services. For example, it may be that parents' work environment may contribute positively to their emerging acculturation to U.S. culture and increasing English language ability. Also, parents who are employed outside the home and use child care benefit from interactions with staff at child care centers by understanding more about child development and U.S. school systems. However, working parents likely experience barriers associated with their work schedules that may impact their willingness or ability to access health care services or educational programs. These issues represent areas of inquiry for future study. --- Limitations Despite the strengths of our study, we must acknowledge several limitations. We measured barriers to treatment just before treatment implementation; therefore, this measure reflected the barriers that parents expected to face in participating in the intervention rather than the actual barriers during the 12 weeks. It may be that a decrease or increase in actual barriers during the intervention is a better predictor of engagement; this should be addressed in future analyses as prior studies ( Mendez et al., 2009;Spoth & Redmond, 2000) assess barriers at one time point treating them as fixed variables. Considering the timing of intervention delivery (e.g., allowing families to sign up for a course during fall or spring) might also make a program more attractive to families, especially when modifiable barriers are anticipated and considered in the intervention planning. Assessing barriers and discussing family concerns is another way to engage families, particularly if other resources (e.g., transportation, child care) can be offered to families before starting program delivery. Another potential limitation of the study involves our sample and associated characteristics. Because our participants were low-income, Latino parents of young children who were U.S. born citizens and who were enrolled in Head Start programs, we cannot state if the findings will generalize to other groups. For example, studies might examine how this treatment might be implemented with immigrants in different settings outside of Head Start or who have lived longer in the United States. Culturally adapted treatments might also impact families differently if children were also immigrants and did not have U.S. citizenship and associated benefits. We also concur with colleagues ( White et al., 2009) who recommend attention to potential differences associated with the culture of origin for Latino populations, and this study was not designed to address this issue. --- Conclusion In conclusion, more research with Latino samples that examines how acculturation may play a role in the selection and usage of health care services is needed. Specifically, studies that consider how acculturation processes unfold over time may help inform what preventive intervention strategies can most effectively support members of immigrant families. Finally, considering what topics are needed for parents with children at different developmental stages is an important aspect of social validity. Auerbach's (2004) Family and Futures intervention is an excellent example of how involving Latino parents in a college access program needs to begin by working with parents of middle school children, to fully engage parents in discussions of college access. The present study reported on a program targeted toward preschool parents addressing the stage-salient issue of school readiness, emphasizing how programming offered in the native language and in a community location may foster treatment engagement among Latinos. Overall, the findings from this study and similar studies regarding efforts to implement culturally adapted treatment for Latino families will help to remedy persistent disparities affecting Latino populations in the U.S. regarding inadequate access to high quality health care and educational services.

Literature to date has not yet included a comprehensive examination of barriers to participation in treatment for Latino populations, incorporating the role of acculturation. This study reports on the process of developing and implementing a culturally adapted treatment for promoting parent involvement by Latino parents in Head Start preschool programs and examines engagement factors affecting participation in the treatment. Results show that Latino parents with higher native cultural competence perceived more benefits to the treatment and reported fewer barriers than parents with lower native cultural competence. Also, the total number of barriers to participation that parents perceived before treatment was negatively associated with treatment participation, above and beyond the influence of acculturation factors, parent perceptions of economic stress, and perceived program benefits. Results are discussed in terms of how to establish effective community-based mental health and educational outreach programs to promote engagement with Latino families. The importance of using bidimensional models of acculturation within research designs with Latino populations and ideas for future research are discussed.Latinos are presently the largest ethnic minority group in the United States. According to the 2010 Census, more than 50 million Latinos live in the United States, accounting for 16.3% of the total U.S. population and 23.1% of children under the age of 18 ( U.S. Census Bureau, 2010). The Latino population represents one of the fastest growing ethnic groups in the United States, showing a 30.3% increase from 2000 to 2010, compared with a 10.6% increase for African Americans and a 1.1% increase for non-Hispanic Whites ( Pew Hispanic Center, 2008; U.S. Census Bureau, 2010). As of 2007, more than half of Latino children were classified as "second generation," meaning that they were born in the United States to immigrant parents ( Fry &

INTRODUCTION It is a proven fact that working under certain circumstances makes people more susceptible to stress and it can potentially entail serious health problems such as depression, anxiety, mood disorders and personality disorders, among others. At the same time, it has also been related to insomnia, lack of emotional control, feeding issues, physical deterioration, the lack of both working and everyday life skills and habits [1][2][3][4][5]. One of the environments where the development of mental health issues among its workers is more common is in correctional facilities. Prisons develop a twofold role: on one hand they serve as a means of punishment for those who have committed a crime and on the other, interns are intended to change their attitude and thereby their thinking by accepting the power of discipline and by assuming the need for submission in order to achieve social rehabilitation, as with psychosocial intervention programs as an alternative to traditional prisons 6. This entails a system where differences are established between interns according to the need for security measures implied by their custody. The work performed in modules together with the roles therein developed -surveillance and guarding of interns, re-education and social rehabilitation as well as other administrative and bureaucratic activities such as processing of penitentiary and criminal files, or health care itself-entail a series of consequences which can potentially influence the health of those performing such tasks 3,[7][8]. To confront this situation an alternative model has been created: the Therapeutic and Educational Unit or therapeutic module. This intends to prioritize on the rehabilitation of the individual as a person as to achieve full social rehabilitation 7,9. Such model intends to confront and end prison subculture, to modify the sole objective of guarding to include the objectives of educating and rehabilitating inmates. Thereby the role of prison workers is also modified and it should be expected that this has a positive influence on both burnout levels and mental health issues among these workers 3,8,10. Rev Esp Sanid Penit 2015; 17: 67-73 C Bringas-Molleda, A Fernández-Mu<unk>iz, E <unk>lvarez-Fresno, A Mart<unk>nez-Cordero, FJ Rodr<unk>guez-D<unk>az Influence of burnout on the health of prison workers Dowden and Tellier 11 assessed work-related stress levels among correctional officers and concluded that certain aspects such as professional commitment or decision-making are highly related to stress and burnout in this context. At the same time, Lambert et al 12, point out that stress among correctional officers negatively influences the development of their work and their own mental and physical health. That is to say that the less involvement in decision-making or the lower the autonomy in the development of their tasks, the higher stress levels tend to be 13. Therefore, if we consider correctional facilities inclusive institutions whose main objective lies in the rehabilitation of inmates and in charge of exhaustively controlling and guarding its "inhabitants", we ask ourselves to what extent burnout entails mental health consequences for those developing their tasks within this environment. Hence our objective is to establish the influence of burnout factors on the health of a sample of prison workers. --- MATERIAL AND METHODS --- Participants This study lies within quantitative explorative descriptive interpretative studies carried out within the institutional context of correctional facilities. The sample has been selected by means of non-probabilistic sampling through volunteer participants from over 10 facilities included in the study to which questionnaires were sent. Hence, the sample is made up of 222 professionals, divided into two different groups: 101 workers from therapeutic modules (which accounts for 45.5% of the sample) and 121 from conventional modules (54.5% of the whole sample). --- Evaluation tools and measures In the first place, evaluation was carried out by means of an ad hoc questionnaire which included personal items such as age, gender, marital status, educational attainment, monthly income, job category, perception of social status, working experience, etc. Then, participants completed Goldberg's and Hillier's General Health Questionnaire (GHQ-28) 14. It includes 28 items on physical and psychological issues, divided into four different aspects. The answers depend on the perception of each individual on their health throughout the last month. There are four possible answers which represent to a greater or a lesser extent the degree of health. The aspects included in this questionnaire as well as the alpha coefficient ob-tained for each of them are the following: somatic symptoms (.898), anxiety/insomnia (.916), social impairment (.764) and depression (.948). Internal consistency of the scale as a whole is.932. Furthermore, these four aspects can be classified in two different groups: Health, with an internal consistency of.786 and Disease,.920. Last, participants were offered Seisdedos' adaptation 16 of Maslach and Jackson's burnout Inventory 15. This questionnaire measures burnout by means of 22 different items, regarding the attitudes of participants towards their job, with answers ranging from 0 (never) to 6 (every day). Furthermore, these 22 items can be classified in three different aspects (all of them with an appropriate degree of reliability): emotional exhaustion -feelings of being emotionally overextended and exhausted by one's work (.875); depersonalization-negative attitudes and passive response towards recipients of one's work (.656) and personal accomplishment-feelings of competence and successful achievement in one's work (.853). The alpha coefficient of the scale as a whole is.882. The last aspect (personal accomplishment) reflects a positive attitude and is therefore related negatively with emotional exhaustion and depersonalization, both of which represent to a certain extent greater levels of burnout. --- Data analysis Statistical analysis of all data was carried out by means of Student's T test, as to assess the existence of significant differences regarding the levels of burnout and health-related aspects between correctional modules-therapeutic and conventional. Later, in order to assess the role that burnout syndrome plays in each level of health-disease, a linear regression analysis was carried out in each group: therapeutic and conventional modules. --- RESULTS The majority of participants were male (N = 164; 73.9%) while the remaining 58 were female (45.5%), age ranged between 18 and 60 years old. Table 1 shows the mean and standard deviation of each of the aspects included in burnout syndrome, the general scale and the variables regarding health both in therapeutic and conventional units as well as in whole. Contrasted means as shown in Table 2 depict significant differences regarding burnout factors between therapeutic and conventional units. Therefore we can see how emotional exhaustion and depersona-C Bringas-Molleda, A Fernández-Mu<unk>iz, E <unk>lvarez-Fresno, A Mart<unk>nez-Cordero, FJ Rodr<unk>guez-D<unk>az Influence of burnout on the health of prison workers lization are significantly higher in conventional modules, although the effect size is small for emotional exhaustion and medium for depersonalization. On the other hand, personal accomplishment is proportionately higher in therapeutic modules with a large effect size. Overall burnout syndrome is more common in non-therapeutic units, with medium effect size. Dimensions of health however, show statistically significant differences as far as social impairment is concerned, which is more common in conventional modules, and depression which shows higher rates in therapeutic units. Nevertheless, a small effect size has been concluded for both, while those referred to somatic symptoms and anxiety/insomnia do not show significant differences. Regression analysis to assess the influence of different burnout categories on health/disease symptoms has allowed the conclusion of certain facts. Hence, as far as somatic symptoms are concerned, Table 3 depicts that within therapeutic modules emotional exhaustion and personal accomplishment play a role in the development of such symptoms. On the other hand, in conventional modules, only emotional exhaustion does so (Table 3). With regard to anxiety/insomnia symptoms, Table 4 depicts how emotional exhaustion is a predictive factor for their development in therapeutic units, while in conventional ones it does so together with depersonalization. As far as social impairment is concerned, again emotional exhaustion plays a leading role in the development of a symptom of health/disease both in therapeutic and conventional modules. Furthermore, depersonalization is also a predictive variable for social impairment in therapeutic modules (see Table 5). To conclude, as far as the last symptom is considered: depression, we have been able to conclude, as depicted on Table 6, that only emotional exhaustion influences most probably the development of depression among prison professionals, regardless of the type of module where they exercise their functions. --- DISCUSSION As shown throughout the analysis of our results, variables that compound burnout syndrome are a consequence of one's working conditions and -69 - the demands and implications entailed by one's job. Furthermore, we know that prisons are one of the working contexts where this syndrome is more likely to be developed 3,8,[11][12]. In addition to this, we have also included the variable health and assessed whether burnout syndrome-related factors negatively affect the deterioration of health. As far as the descriptive analysis of the variable burnout and Health is considered in each of the intervention alternatives of the prison context (therapeutic and conventional units), we have observed that the variable burnout in general has a broader impact in conventional modules. When considering its dimensions we have determined that Emotional Exhaustion and Depersonalization are more common in workers from conventional units (non-therapeutic) and that Personal Accomplishment prevails in the alternative perspective that therapeutic units represent. Such differences are in line with what Bringas et al 13 determined establishing disparities regarding the correlation of Personal Accomplishment and the perception of a negative or positive working climate. Higher incidence rates of burnout syndrome among workers from conventional modules may be due to further exhaustion and occupational stress, with a lack of interest towards work and a series of negative attitudes and behaviours towards recipients of one's work, who are sometimes not easy to cope with-something which occasionally entails feelings of guilt over unsuccessful social relations between inmates and officers. These results are also in line with the conclusions by Hernandez 13, who found high levels of exhaustion and depersonalization among officers and low rates of personal accomplishment with low expectations. With reference to the dimensions of the variable Health, Social Impairment prevails (without distinguishing between conventional and therapeutic modules) although there are no significant differences with the other three dimensions. Social relations are one of the basic pillars of this profession and obviously balanced and healthy relations are difficult to keep in the prison working environment-a fact that would explain in part a certain degree of social inhibition and withdrawal. On the other hand, we have observed that Depression plays a more significant role in therapeutic modules, while Social Impairment prevails in conventional modules. We have determined that Depression is the dimension of Health whose development is less probable in both modules. When considering the influence of burnout-variables over Somatic Symptoms, we have observed that Emotional Exhaustion and Personal Accomplishment prevail as potential agents for the development of so-matic symptoms among workers from therapeutic units. It would appear that people under the effects of stress experiment further physiological activation which entails further awareness of such states and a higher probability of labeling symptoms as signs of disease. This makes sense, since emotional exhaustion and weakening entail difficulties handling stress, a lack of control, unpredictability and other common factors. In the study by Arroyo et al 17, no significant differences were concluded for the development of hypochondriasis in prison, a fact which could mean that it is similarly found in other working contexts although it could also be explained by the consideration of other variables such as Personality and Perception of the working climate, which could have influenced the results. Presumably, Personal Accomplishment influences the development of these symptoms that suggest the appearance of a physical disorder without a clear organic cause or physiological mechanism, and so they may be associated to stress factors. We say presumably because it has not been proven that self-fulfillment, self-esteem and a positive perception of personal accomplishment entail the presence of hypochondriasis. The questionnaire used is an identification method of patients designed for outpatient consultation and to detect alterations of standard function, which may mean that they are not necessarily ill. With respect to conventional modules, we have concluded that the most influential variable is Emotional Exhaustion. In other studies which assess workers from correctional facilities [18][19], high rates if physical and emotional exhaustion were found among such, as well as higher rates of burnout syndrome in comparison with workers from other settings. As far as the probability of developing anxiety and insomnia, the variable emotional Exhaustion scores higher in therapeutic modules. Sleep disorders are mostly related with a series of emotional disturbances which can, in many cases, appear as anxiety symptoms and hence both factors are consistent between them. Indeed, these features can be a consequence of Emotional Exhaustion as the influence negatively in success expectations and efficiency, which are also related to self-esteem and can imply the appearance of anxiety symptoms and sleep disorders. In conventional modules, Emotional Exhaustion and Depersonalization are the main predictors for the development of Anxiety and Insomnia. Adherence to the later in alternative therapeutic modules can be due to further personal disrealization and its consequent negative attitudes in the relations with inmates due to constant surveillance and alert functions. The attitudes derived from depersonalization can entail states of nervous anxiety and sleep disorders since they imply high levels of exigency and continuous pressure. Another factor included in the variable Health is Social Impairment. In this case we find the opposite situation: Emotional Exhaustion and Depersonalization act in the therapeutic modules and only Emotional Exhaustion in conventional ones. Social impairment is developed after low levels of wellness and a negative perception of social skills implied in the development of one's work, which mostly influence social relations. According to Lambert et al 12 stress, understood as working and emotional exhaustion, influences physical and mental health, while it is also true that emotional exhaustion entailed by working stress together with depersonalization, with negative attitudes towards the rest, may lead to Social Impairment both in the alternative and conventional prison perspectives since there are no significant differences in this regard. Depression includes symptoms such as indifference, impaired concentration, dejection, low self-esteem and a sense of futility, among others. According to our results this can be due to Emotional Exhaustion, without significant differences between modules. Again emotional and working weakening, due to the pressure entailed by such works, seem to be responsible in the development of the aforementioned feelings, at least among prison officers. Pressure derived from coping with inmates and keeping a state of alert throughout most of their working time implies emotional exhaustion and this leads to a decay of performance, positive perception of one and one's work and increased feelings of sadness or melancholy. This study has focused on the analysis of the influence of burnout categories over different healthrelated factors. We believe that it would be necessary to assess the previous health status of workers, at least one month before, as established by the test GHQ-28. Previous health status may play a role in the development of one's tasks and therefore levels of emotional exhaustion and/or depersonalization may influence one's health status. Likewise, it is advisable to assess the role that each health factor plays in the development of burnout syndrome. In conclusion, burnout syndrome has consequences on the deterioration of health. Its dimension can act as predictors in the development of disease in the context of correctional facilities. Nevertheless, such effects are not exclusive, since such deterioration can be developed regardless and previous to burnout syndrome. On the other hand, it would be necessary to consider potential bias of this study, such as the level of prison officers, the type of tasks developed in a se-ries of aspects, for example to what extent they keep a relation with inmates, which can affect the different dimensions included in the burnout Syndrome and in the variable Health. --- CORRESPONDENCE Francisco Javier Rodr<unk>guez D<unk>az Facultad de Psicolog<unk>a. (School of Psychology) Despacho 215 (Office 215) OVIEDO (Spain) [email protected]

Objective: This study investigates differences in health (Somatic Symptoms, Anxiety / Insomnia, Social Dysfunction and Severe Depression) relating to professional burnout (emotional exhaustion, depersonalization and personal fulfillment) suffered by prison staff, taking into account the workplace where they perform their activities (therapeutic vs. non therapeutic modules). Methods: The participants consisted of 222 workers of both sexes, with ages ranging from 18 to 60 years, working in different prisons in Spain (Educational and Therapeutic Modules, N = 1001; Non Educational and Therapeutic Modules = 121). The tools used were the General Health Questionnaire (GHQ-28) of Goldberg and Hillier (1979) and the burnout Inventory of Maslach and Jackson (1981). A mean comparison was performed using Student's t test along with a linear regression analysis, differentiating between Educational and Therapeutic Modules and Non Educational and Therapeutic Modules. Results: Significant differences were found between both clusters and predictive values of burnout factors for the different health levels in both prison models were identified. Possible explanations for the findings and the implications for intervention are discussed.

Background Two of the most distinctive features of overall mortality in Finland are that death rates increase in the southwest to northeast direction and that life expectancy of Finnish speakers is shorter than that of Swedish speakers [1]. Both these aspects conform to recent genetic mappings [2,3]. There are differences in the genetic architecture that increase in the southwest to northeast direction, and Swedish speakers in Finland have been found genetically closer to Swedes than to Finnish speakers. The variation owes to the country's settlement history. Finns are unique on the genetic map of Europe, because they were at one time a small and geographically isolated population. Being born in a high-mortality area in Finland is associated with a 10-30 per cent higher risk of dying from ischemic heart disease at high ages, as compared with being born in a low-mortality area [4]. A similar geographical pattern dominated by people's birth region has been observed for deaths from other common causes and in various age groups [5][6][7]. Hence, hereditary components might lie behind many deaths from cardiovascular diseases as well as external causes such as suicides and alcohol-related deaths. Many diseases, such as coronary heart disease, psychiatric disorders and alcoholism, are nowadays known to be associated with genetic susceptibility [8][9][10]. Taken together, these findings illustrate that, in the particular case of Finland, large-scale register data have the potential of shedding light on the question of how human genetic diversity manifests in mortality risks at the population level, even if they lack explicit genetic information and biomarkers. The country's settlement history consequently provides a rationale for interpreting the regional mortality differences as reflecting geographic clustering of hereditary factors. No explicit link between the mortality variation and genetic differences has yet been established, however. All previous studies on the effect of geographic ancestry on mortality risks in Finland have studied the influence of persons' own birthplace as compared to that of their current place of residence. That approach for reflecting geographic origin may be contaminated by the past decades' internal migration flows. The birthplace of a parent, on the other hand, is a substantially better indicator of geographic family origin, as it widens the time horizon and is not confounded by person-specific environmental or behavioural factors. This paper constitutes the first large-scale study of the Finnish regional mortality differences with focus on the influence of parental birth area. The opportunity was provided by recently available intergenerational data from longitudinal population registers. Assuming that geographic family origin reflects hereditary factors that influence mortality, we expected to find an effect of parental birth area on death risks irrespective of where the study persons were born or living as adults. Hence, even if people were born and raised in a low-mortality area, they would have an elevated death risk if a parent was born in a high-mortality area. --- Methods The data came from the longitudinal population registers of Statistics Finland and covered the period 1970-2003. The part used here consists of an eight per cent sample of all Finnish speakers, and an additional 50% sample of all Swedish speakers. The latter constitute a national minority of barely six per cent of the total population. Basic demographic and socioeconomic variables are available for every fifth year, based on the quinquennial population censuses. Deaths from all causes and emigration are known for each year. For all persons born after 1953, there is corresponding information for the parents. The structure implies that individuals could be observed at most up to the year when they become 50 years. Since there are few deaths at childhood, adolescence, and young adulthood, we started observing individuals from the year in which they became 30 years. The analyses consequently concerned ages 30-49 years, the period 1985-2003, and people born 1953-1970. Oneyear death risks were estimated with Cox regressions using age as the duration variable. Figure 1 illustrates the southwest to northeast increase in mortality rates standardised for age and calendar year, based on official aggregate statistics comprising the total male population by region for the period 1986-2005. The regions have been geographically numbered and coloured to facilitate comparisons. The thick broken line represents the most evident mortality difference, which roughly corresponds with the first national boundary between Sweden-Finland and Russia in 1323. A similar divide has been observed in terms of anthropology, folklore, and dialects [11]. In the data, there is unfortunately no information about the cause of death. However, from previous research we know that the regional mortality variation is dominated by mortality from cardiovascular diseases, alcohol-related causes and suicides [5], which also are the most common main causes of death in men aged 30-49 years (Table 1, which is based on [12]). Mortality in same-aged women is outside the scope of this paper, as it is modest and dominated by various non-vascular diseases that display no evident regional pattern. To obtain enough statistical power for the analyses, a simplified regional categorisation was used. It distinguished two groups of Finnish speakers according to the east-west boundary mentioned above. For each region variable of interest -the person's area of residence (at age 30-34 years), the person's birth area, mother's birth area and father's birth area -the same categorisation was applied. A separate category was applied for the Swedish speakers, independent of their own birth area or parental birth area, because most of them originate in western Finland. As much as 97% of all Swedish speakers live in the regions that are white-coloured on the map, and only two per cent in other parts of the western area. The remaining one per cent was excluded from analysis. We accounted for effects of some central personal characteristics in terms of each individual's birth year, educational level (primary, secondary, lower-degree tertiary, or higher-degree tertiary) and family type at childhood (one or two parents in the household at age 10-14 years). The latter was known because of the longitudinal structure of the data, meaning that all persons could be observed at the time when they were aged 10-14 years. All these variables are important mortality determinants in Finland [5][6][7]. There was information also about parental education and parental socioeconomic position. These variables were excluded from the models presented here, because they correlated strongly with the study persons' educational level, but had no impact on the estimates for the other variables. --- Results The unweighted data comprised 61531 persons, whereof 47434 were Finnish speakers and 14097 were Swedish speakers (Table 2). The total number of deaths was 1464. The importance of internal migration on the population structure can be seen from the distribution of Finnish speakers on the region variables. Approximately 60% of the Finnish speakers were born in western Finland, but almost two thirds had at least one parent from eastern Finland. Even among those who were both born and residing in western Finland, the share with at least one parent born in eastern Finland was as high as 40% (see next section). Two thirds of the birth cohorts were living in western Finland at age 30 years. In a first model, we included persons' area of residence and the control variables. The results, which are summarised in Table 3 highlight some known facts. People residing in eastern Finland had 12% higher death risks (95% confidence interval 1.00 to 1.26) than those residing in western Finland, whereas the death risk of Swedish Finland is equal to one. The classification is according to the 20 administrative regions, plus one category (number 7) that separates the Helsinki metropolitan area. Authors' calculations based on [12]. speakers was 40% lower. In these ages, the ethnic-group mortality difference is the widest [6]. Also the control variables had substantial effects. Mortality risks decreased notably over educational levels and birth cohorts, whereas people raised in one-parent families had almost 40% higher death risks than those originating in two-parent households. The inclusion of these variables did not reduce the area or ethnic mortality difference, however. We proceeded by focusing on the influence of each area variable (Table 4). The control variables were also included, but since the estimated effects were the same as reported above, they are not shown here. Replacing area of residence (model 1) with own birth area (model 2) or mother's birth area (model 3) slightly improved the Wald statistic and increased the area difference marginally. Father's birth area (model 4) had a stronger effect and improved the statistical fit relatively more. People with the father born in eastern Finland had 21% higher death risks (1.08 to 1.35) than those with the father born in western Finland. When using either mother's birth area or father's birth area, the reference category is heterogeneous with respect to the study persons' geographic family origin, because the other parent may be born in eastern Finland. Simultaneously accounting for both parents' birth area (model 5) revealed that the offspring death risk was 23% higher (1.09 to 1.39) if at least one parent was born in eastern Finland, as compared with people whose both parents were born in western Finland. Having both parents from eastern Finland was associated with only a slightly higher mortality risk as compared with having only one parent from eastern Finland, however (model 6). The importance of parental birth area can also be illustrated by adding it into either model 1 or model 2. The effects of area of residence and own birth area then totally disappear (Table 5), suggesting a marginal role of areaspecific current or early-life factors. Having a parent born in eastern Finland was still associated with more than 20% higher death risks as compared with having both parents born in western Finland. Another way to contrast the influence of geographic family origin against that of current and childhood environmental factors is to restrict the study group to people born and residing in western Finland. Doing so, we found that persons with at least one parent born in eastern Finland had 21% higher death risks (1.04 to 1.41) than those whose both parents were born in western Finland (Table 6). Hence, there was only a modest reduction in the effect of parental birth area as compared to that of model 5 in table 4. --- Discussion and conclusions Our results correspond with known facts about mortality of middle-aged men in Finland. The death risk of Finnish speakers in western Finland is approximately ten per cent lower than that of Finnish speakers in eastern Finland, but 40% higher than that of Swedish speakers. These mortality differences resemble genetic mappings of the population, but no explicit link has yet been established. As related to previous research based on population registers, which has documented mortality effects of people's own birth area and area of residence, we have here proceeded by using intergenerational data and focusing on parental birth area. We find that mortality differences by people's own birth area, which might proxy not only genetic factors but also early-life conditions, are fully explained by the parents' birth area. If at least one parent was born in eastern Finland, the death risk is over 20% higher as compared with if both parents were born in western Finland. A similar elevation is the case even if we restrict the study group to people born and residing in western Finland. Hence, geographic clustering of hereditary factors, as assumed reflected by parental birth region, seems to manifest in increased mortality risks. A limitation of the study is that we have no information about specific causes of death. However, we do know that in the ages studied, the dominating causes of death may, at least to some extent, be genetically related. It also needs to be stressed that, since we cannot trace any other ancestors than the parents, parental birth area as used here is most likely an underestimate of the true effect of geographic ancestry on mortality. Since internal migration in modern time has been directed primarily towards the western parts of the country, the current population in this area is genetically mixed. Many persons with both parents born in western Finland are therefore likely heterogeneous with respect to the birth area of second or higher order generations. Identification of family origin several generations back in time would accentuate any hereditable component predisposing to complex disorders [13]. An alternative explanation to the sizeable effect of parental birth area on offspring mortality would be learned patterns of behaviour. In that case, parents' poor diet and inadequate coping with risks in life would be transferred to the next generation in a patriarchal manner. The role of such social inheritance, as opposed to that of genetic factors, cannot, in our opinion, be large. It would imply that, even if the children were born, raised and living in a low-mortality area, they have embraced parents' supposed unhealthy behaviours. It also postulates that there would be some latent factor associated with the eastern Finnish way of life that is highly detrimental for health, and independent of whether it is the mother or the father that carries it. In families where one parent is from eastern Finland and the other from western Finland, it must also dominate any socially protective effects from the parent born in the low-mortality area. Various mappings of the Finnish population's health behaviours have been undertaken during the past decades. An overall conclusion based on these surveys is that dietary practices and culturally determined lifestyle factors do not underlie the regional mortality differences [14]. On the contrary, there is ample evidence about substantial genetic diversity across regions, and of susceptibility to specific diseases according to people's geographic roots [4]. During the past decade, the annual number of deaths in middle-aged men in eastern Finland has been approximately 700. As compared to same-aged Finnish-speaking men in western Finland, the excess mortality is more than 100 deaths per year. As the geographic mortality divide is strongly present also at higher ages [7][8][9], the total effect is even larger above the middle ages. At ages 50-74 years, for instance, the annual number of deaths by cardiovascular diseases, alcohol-related causes, and suicides in eastern Finland is almost 2000. This corresponds to an excess mortality of more than 300 deaths as compared to western Finland. From a public health perspective, many early deaths may consequently be avoided if genetic predisposition is recognised via close monitoring of persons' family origin. --- Authors' contributions Both authors contributed equally in conceiving and designing the study, analysing and interpreting the data, and preparing the manuscript. --- Competing interests The authors declare that they have no competing interests. --- Pre-publication history The pre-publication history for this paper can be accessed here: http://www.biomedcentral.

Background: Death rates are notably higher in eastern Finland than in western Finland, and life expectancy of Finnish speakers shorter than that of Swedish speakers. The mortality differences correspond to recent genetic mappings of the population and are prominent for causes of death that are known to be associated with genetic risk factors. Methods: Using intergenerational data, we studied the impact of parental birth area on all-cause mortality risks of middle-aged men in Finland 1985-2003, assuming that geographic family origin reflects genetic predisposition to complex disorders. Relative death risks at ages 30-49 years were estimated by parental birth region and ethnicity, according to Cox regressions standardised for own education, family type at childhood, and year of birth. Results: The death risk of Finnish speakers born in eastern Finland was 1.13 (95% confidence interval 1.01 to 1.26) that of Finnish speakers born in western Finland, whereas that of Swedish speakers was only 0.60 (0.52 to 0.71). In Finnish speakers, the effects of own birth area and area of residence disappeared when parental birth area was accounted for. The death risk of persons with at least one parent born in eastern Finland was 1.23 (1.09 to 1.39) that of people with both parents born in western Finland. Conclusions: Parental birth area is the driving force behind the regional mortality difference in Finland. The findings highlight and give further support for the potentially important role of genetic risk factors in mortality. Close monitoring of persons' geographic and ethnic ancestry may promote public health and avoid many early deaths.

Introduction According to Article 5 of the Universal Declaration of Human Rights, ''no one shall be subjected to torture or to cruel, inhuman or degrading treatment or punishment.'' [1]. This paper presents evidence of continued female genital mutilation (FGM) and gavage (force-feeding of adolescent females) in Mauritania. While these practices are culturally sanctioned in the Mauritanian context, the harming of genital organs of any child (including harm resulting from the practice of gavage and FGM) is illegal under the child protection penal code; penalties range from 1-3 years imprisonment and heavy fines. However, the law does not specify FGM or gavage as explicit illegal practices, although both practices have the potential to result in pain and psychological suffering on the girl. Further, FGM has been linked to the development of feelings of incompleteness, loss of self-esteem, depression, chronic anxiety, phobias, panic and psychotic disorders [2]. No information was found on the psychological effects of gavage; this void in the literature represents an important area for further investigation in Mauritania. Whereas FGM has been documented in numerous sub-Saharan Africa (SSA) countries [3][4][5][6] and is widely recognized by the international community as a human rights violation and public health risk to women in countries where it is practiced [2], gavage is somewhat unique to a few African countries and relatively unknown to many, including experienced international public health professionals. Given that gavage has been linked to child marriage and violence against young girls [7][8][9], and both FGM and gavage are a public health concerns and violate the rights and welfare of children, more needs to be done to understand the practices before effective challenges to these cultural norms are launched. --- Female Genital Mutilation FGM, also known as female genital cutting describes all procedures that involve complete or partial removal of the external female genitalia, or injury to the female genital organs for non-therapeutic reason. It has no known health benefits and may harm girls and women permanently [3,[10][11][12][13][14][15]. There are four general types of FGM practiced [14], although the type and extent varies by country and culture: 1) Clitoridectomy, where the clitoris is either partially or totally removed; 2) Excision, where the clitoris and labia are either partially or totally removed; 3) Infibulation, which refers to a condition where the vaginal orifice is narrowed by a seal created by cutting and transposing the labia majora or the labia minora; and 4) other practices that include cauterizing, incising, pricking, scraping or piercing the genital areas. FGM is generally believed to curb a woman's sexual desire and many assume it is required for religious reasons; it is also done to confer social acceptance on the family, as well as a rite of passage from childhood to adulthood [16][17][18][19][20]. The majority of FGM practiced in Mauritania are types I and II; FGM is typically done at a very early age ranging from a few weeks old to 3 years of age. --- Gavage Gavage is a French term that literally means ''force feeding''; force-feeding is a method used to fatten geese for foie gras. In Mauritania, gavage is known as ''Leblouh''. The practice of fattening girls has potentially harmful consequences related to reduced movement and increased risk of developing cardiovascular diseases [8]. According to the World Health Organization (WHO), approximately three million adult deaths occur every year resulting from obesity. As well, a significant proportion of diabetes, heart diseases and cancers are attributed to obesity or being overweight. [21]. Gavage is predominantly practiced among Arab families in Mauritania; while there are also examples of pre-marriage ''fattening houses'' for women among other ethnic groups in West Africa, this fattening is typically done at a later age, is done only for the purpose of the immediate marriage, and is usually consensual. Mauritania ranks number one on a list of countries that celebrate female obesity; this celebration poses human rights violations if girls are being forced against their will to fatten, in addition to the public health concerns related to obesity. A second problem relates to a lack of financing to combat the negative physical and psychological health effects of obesity in Mauritania [22][23]. As the social norms for marriage favor young girls that have a large amount of body fat, families often force young girls (sometimes as early as 6 years old) to consume large quantities of food and milk to increase their weight, and thus their chances for marriage at an early age. The use of pills (e.g. steroids) to gain weight has also been reported in Mauritania. Thin women are believed to be unhealthy, whereas overweight women are perceived as far more desirable [24][25]. Further, the large size of these force-fed girls creates an illusion that they are physically mature and ready for marriage; this creates the potential for reproductive health and psychological problems later in life if the girl has not yet reached maturity. There is a little reference as to when force-feeding of girls began in Mauritania. Some historians believe that the practice dates back centuries when most of the white Mauritanians, Arabs, and Berbers were nomads; in nomadic society the obesity of women was seen as a sign of beauty and prosperity and the wives of rich men would often not work, preferring instead to sit in tents while black slaves tended to household chores. [26]. Ostensibly, FGM and gavage have similarities. Both are practiced on girls, and mothers or grandmothers often perpetuate these practices. Both also have strong gender implications relating to harm performed in the name of increasing a girl's desirability for marriage or social acceptance. And, both represent serious human rights violations, as these practices produce physical suffering and have implications for long-term psychological well-being. There is a dearth of published literature or statistics on public health problems in Mauritania in general, and even less on these two practices specifically. This paper constitutes the first published report to examine these two practices in terms of prevalence, prevailing attitudes and acceptance, and factors associated with their use in Mauritania. While the data used are almost a decade old, the 2001 demographic and health (DHS) survey constitutes the only known population-based survey to collect data on both FGM and gavage. The existence of two such harmful practices to girls in a single country highlights the importance of surveillance and investigations into these practices simultaneously. As such, the objectives of this paper are 1) to document the prevalence of FGM and gavage in Mauritania, 2) to investigate gender differences in attitude towards these practices, and 3) identify factors associated with FGM and gavage practices. The results presented herein are intended to increase awareness and serve as a platform for developing culturally relevant prevention mechanisms in Mauritania and elsewhere with the long-term goal of improving the reproductive and psychological health of girls. --- Methods --- Study Site Mauritania is a country located in Northwest Africa between Morocco, Western Sahara and Senegal. It is one of the poorest and least developed countries in the world. In 2011, the country's Human Development Index (HDI) was very low (0.453), ranking 159 th out of 182 countries [27]. Over half (58.6%) of the population lives in rural areas; the dispersed settlement pattern often poses health care access challenges for many communities. Of the country's four ethnic groups (Arab, Poular, Soninke, and Wolof), Arabs are by far the largest. [8]. Life expectancy is 56 years for males and 60 years for females. The maternal mortality rate is 550 deaths per 100,000 live births. Only 4% of the central government expenditure is allocated to health, representing 2.2% of gross domestic product [23]. Mauritania ranks 114 th out of 135 countries on the Gender Gap index [28]. --- Study Design and Sample Size Data were obtained from the Mauritania 2001 Demographic and Health survey (DHS) [8]. This was a secondary analysis of DHS data; the Mauritanian government gave permission to use these data. The Tulane University Institutional Review Board (IRB) reviewed and approved the analysis only, as Tulane University researchers were not involved with data collection or study design. The DHS survey was approved by the Mauritanian Government (National Direction of Statistics). The DHS targets women of reproductive age and therefore a statement about children is not needed. Due to high illiteracy rates in the area, verbal consent and assent, in lieu of written consent and assent, were obtained during the collection of data. No data were collected from respondents who did not give permission to participate. All interviews were conducted in private to reduce social desirability bias. All data were collected using a standardized structured questionnaire [8]. The survey used a two-stage cluster design to produce a probability sample, representative of the entire country; data were collected on women of reproductive age (15 to 49 years old), and men (15 to 59 years old). A total of 6,149 households were successfully surveyed; only 1.3% of households approached refused to participate. Women between the ages of 15 and 49 years old in selected households were interviewed individually; a total of 7,728 were interviewed, with 3.5% of eligible women refusing to participate. Forty-percent (40%) of households were selected for male interviews; a total of 2,191 men between the ages of 15 and 59 years old were interviewed, with 11.9% of eligible men refusing to participate. --- Data Analysis STATA 9.0 (STATA Corporation, College Station, TX) was used for all data analyses. Descriptive statistics were used to summarize 2001 DHS household survey data by demographic characteristics. Chi-square test statistics were used to explore differences between demographic characteristics and the following 6 outcomes: attitude about continuing the practice of FGM (among men and women separately), attitude towards continuing the practice of gavage (among women and men separately), the prevalence of FGM among women, and the prevalence of gavage among women. Individual level demographic characteristics included age and education. Household level demographic characteristics included wealth, as measured by a household asset index created using a principal components analysis of household assets (e.g. electricity, radio, television, refrigerator, vehicle and animals) [29]; place of residence (urban versus rural and region); and ethnicity (Note: in some analyses, ethnicity categories were either dropped or collapsed to increase the sample size within categories). Lifestyle binary variables included marital status and working status. While we recognize that other variables may also be important, the socioeconomic and demographic variables described above were chosen as a starting point for investigations. Logistic regression models were used to test whether sociodemographic factors were associated with the 6 outcomes related to attitudes and practices of FGM and gavage. Regressions using FGM outcomes excluded region of residence due to collinearlity with residence in rural versus urban areas. An interaction term between place of residence and working status was included in both of the multivariate models; no other interaction terms tested significant. Regressions on attitude were done among respondents who had heard of FGM or gavage. A second set of logistic regression models were used to test for factors associated with the prevalence of both FGM and gavage among women who have heard of the respective practices. All regressions included only those respondents with complete data and reported knowing of the respective practices. No differences were detected in prevalence or attitudes between those with complete data records and those within incomplete data records. To control for intra-class correlation at the community level, empirically estimated standard errors were used within logistic regression models, using the village as the cluster unit. Probability weights were used to account for the two-stage cluster sample design used. The probability of committing a type-1 error (alpha) was set at 0.05. Wald statistics and log-likelihood ratios were used to identify variable significance and model fit. --- Results Table 1 presents data on the socio-demographic profile of the sample. As expected, approximately 75% and 20% of the sample were Arab and Poular, respectively. About a quarter of women were single at the time of the survey, compared to almost half of men. Roughly half of the respondents were from urban areas, with almost 40% living in the capital city of Nouakchott. --- Knowledge, Attitude, and Prevalence of FGM In Mauritania, over 70% of women report experiencing FGM [8]. While the extent of this practice varies somewhat as a function of cultural practices and location in Mauritania, it has been reported as widespread [8,15]. perceived to be the advantage of FGM, the most common responses were higher social recognition, fulfilling a religious requirement, and reducing female sexual desire. Over half of both male and female respondents reported that there are no disadvantages to the practice of FGM (53% of women and 58% of men). The majority of both female and male respondents favored the continuation of the practice (64% and 70%, respectively). Interestingly, there was discordance between male and female beliefs that that opposite sex desired the continuation of the practice, with 37% of women reporting that they thought that males wanted to continue the practice and 55% of men reporting that women wanted to continue the practice of FGM. The practice of FGM among respondents in this sample was widespread, with 77% of women reporting that they experienced FGM. Among respondents (both male and female) who had at least one daughter, 71% reported practicing FGM on their daughter. Bivariate results indicated significant differences in attitude (i.e., approval of the continuation of FGM) for both male and female respondents. The most significant differences were observed among women and among men who live in different regions [female (x 2 = 1,260.6, p,0.001), male (245.9, p,0.001)], with both women and men who live in Nouakchott less likely to approve of the continuation of FGM. Women and men in the highest wealth quintile were also less likely to approve of continued FGM [female (x 2 = 550.1, p,0.001)], male (x 2 103.2, p,0.001)]. Women and men in rural areas were more likely to approve of continued FGM [female (x 2 = 310.5, p,0.001), male (x 2 = 126.9, p,0.001)]. --- Knowledge, Attitude, Intention to Practice, and Prevalence of Gavage In Mauritania, over 20% of women report experiencing gavage [7]. Table 3 presents data on attitudes toward gavage. Almost all respondents had heard of gavage (92%). Approximately a third of respondents reported that this practice increased a girl's beauty (40% of women and 30% of men) and about a quarter reported that the practice increased the family's social standing in the community (27% of women and 21% of men). In contrast, nearly 40% of women and 55% of men reported that gavage has no advantages. The most common reason given to not practicing gavage was improved health. In contrast, approximately a quarter of both men (23%) and women (25%) cited no disadvantages with the practice of gavage. Among respondents with at least one daughter, 17% of women and 11% of men reported the intention to practice gavage. Almost a quarter (23%) of women reported being force fed as a child; among those with a daughter, 6% of women and 11% of men had already force fed at least one daughter. Over 61% of those who had experienced gavage reported being beaten during the process and almost one-third (29%) reported having their fingers broken to encourage participation. Conversely, among those women who regretted being force-fed, over half (58%) reported health risks as the most common reason; over one-third (34%) reported that being big made it too difficult to work and move. Pain was also a common reason (31%) cited for regret. Bivariate results indicated significant differences in attitude (approval of the continuation of gavage) for males and females. The most important differences in attitude toward the continuation of gavage were observed among women and among men from different ethnic groups, with Arabs more likely to approve of the continuation of gavage [female (x 2 = 358.8, p,0.001) male (x 2 = 49.9, p,0.001)]. Men and women from Nouakchott were also more likely to approve of continued gavage: [female (x 2 = 351.2, p,0.001), male (x 2 = 374.0, p,0.001)]. Those from the lowest wealth quintile also supported the continued practice of gavage: [female (x 2 = 257.2, p,0.001), male (x 2 = 111.1, p,0.001)]. --- Multivariate Analysis of Socio-demographic Factors Related to FGM and Gavage Results from the 3 logistic regressions with FGM attitude and practice outcomes are presented in Table 4. The strongest predictors for approving (attitude) of continued FGM were Women between the ages of 15 and 19 were also more likely to report approval, although age was not a significant factor for men. The predictors of FGM approval (noted above) were also important in the model assessing relationships between socio- Results from the 3 logistic regressions with the outcomes of attitude and practice of gavage are presented in Table 5. As with FGM, ethnicity, location, education, wealth, and age were significantly associated with approval of the continuation of gavage. In this analysis, Arab women and men were 4.96 (95% CI: 3.07-8.00) and 2.52 (95% CI: 1.53-4.18) times more likely to approve of the continuation of gavage, respectively, as compared to other ethnicities. Males from Hogh cha were 14.38 (95% CI: 6.60-31.31) times more likely to approve of the continuation of gavage as compared to males residing in Nouakchott. Also noteworthy is that This analysis was done among those women who know (heard of) the practice of FGM; we excluded women who did not know of FGM, did not belong to one of the 4 major ethnic groups, and the missing variables. --- <unk> This analysis was done among those women who either approve or disapprove the continuation of FGM; we excluded women who did not know of FGM, did not belong to one of the 4 major ethnic groups, and the missing variables. doi:10.1371/journal.pone.0060594.t004 any education and high wealth reduces the odds of approving the continued practice of gavage, as compared to no education and low wealth. Older women were less likely [O.R. 0.51 (95% CI: 0.38-0.68)] to approve of the continuation as compared to women between the ages of 15 and 19 years old. One interaction term was determined to be significant, whereby having a job and being located in a rural area almost doubled the odds [O.R. 1.67 (95% CI: 1.22-2.28)] of approving the continued practice of gavage. Results from the regression with the outcome of having experienced gavage were similar to those related to attitude towards gavage. Importantly, Arab women were over 17 (O.R. 17.46 (95% CI: 10.75-28.35)] times more likely to report having experienced gavage, as compared to other ethnicities. While education reduced the odds of approving the continuation of gavage, in this analysis having Koranic education doubled the odds [O.R. 2.12 (95% CI: 1.76-2.59)] of having experienced gavage, as compared to having no education. In this analysis, the interaction term between having a job and being located in a rural area decreased the odds [O.R. 0.69 (95% CI: 0.51-0.93)] of having experienced gavage. This analysis was done among those women who know (heard of) the practice of gavage; we excluded women who did not know of gavage, did not belong to one of the 4 major ethnic groups, and the missing variables. --- <unk> This analysis was done among those women who either approve or disapprove the continuation of gavage; we excluded women who did not know of gavage, did not belong to one of the 4 major ethnic groups, and the missing variables. doi:10.1371/journal.pone.0060594.t005 Overlap between FGM and Gavage An investigation into the potential overlap between practices among women suggests that in general little overlap exists. One in five (19%) reported experiencing both gavage and FGM; 19% reported experiencing neither. Only 4% reported being force-fed, but not cut; 58% reported experiencing FGM, but not gavage. --- Discussion Results from this study suggest that both FGM and gavage are actively practiced and approved of in Mauritania; these practices are unlikely to be discontinued in the near future without considerable effort to change cultural norms. Region-specific interventions should be developed to increase awareness about the public health and psychological dangers these practices pose. Given that Mauritania has already committed to protect the rights and welfare of children [30], this existing framework should be used to challenge these cultural norms. While these practices are illegal if they cause pain or suffering to children, perpetrators are not held accountable, likely the result of engrained tradition that encourages these practices [7,31,32]. It is also possible that by prosecuting offenders the practice would be perceived as a cultural attack and largely driven underground, rendering surveillance and the development of intervention to change cultural norms difficult. The results of this study are important, as they constitute the first study around the practice and attitude of women and men toward FGM and gavage in Mauritania. These results might further serve as a base for developing evidence-based strategies to combat these harmful practices. Observations related to important predictors of FGM and gavage are discussed below, in the context of how best to protect the rights and welfare of children. Both practices appear to be largely tied to ethnicity, although the ethnicities are different for FGM and gavage. FGM for example was highest among the Soninke and lowest among the Wolof, consistent with other regional studies [33]. Gavage however was most prevalent among Arab women. These observations provide a starting point for identifying which groups could benefit from increased awareness and education related to the public health and psychological dangers these practices pose. Awareness raising campaigns focused on behavior change communication could then be integrated into media and traditional information sources within regions most likely to practice the respective traditions. While other studies have produced mixed results on associations between education and FGM, [3,[33][34][35][36][37][38], this analysis showed education to have a protective effect on attitude and practice of both FGM and gavage. Although the practice of gavage was positively associated with Koranic education, this finding is likely explained by the large proportion of Arab women with Koranic education confounding our results. Educators in collaboration with civil authorities should continue their focus on destigmatizing the refusal of these practices. Other interventions could target school-aged children, thus creating a generation with the information to question the merit of these practices. One perplexing finding from this analysis was the seemingly contradictory role of age in relation to attitudes and practice of gavage. Older women were more likely to have experienced gavage than their younger counterparts, which might be expected if the prevalence was even higher in the past. However, positive attitudes toward the continuation of the practice of gavage were higher among younger women than older respondents. One explanation may be that these older women have experienced the health consequences of gavage, such as diabetes or cardiovascular diseases, and have thus received information linking these conditions to obesity; younger women may not have experienced any health consequences associated with obesity, or may not be targeted for chronic disease and healthy lifestyle information dissemination. A second explanation may simply be that women who have experienced gavage recognize that this is torture, and thus less likely to have a positive attitude regardless of the long-term health consequences. It is important that all age groups be targeted for strategic intervention, although the delivery mechanism will likely need to be tailored between age groups to improve effectiveness. A second perplexing finding relates to the discordance between those that reported no disadvantages to the practice of FGM and those that favored its continuation. In this analysis, it was clear that some respondents (both male and female) reported disadvantages related to the practice of FGM yet favored its continuation. While the explanation is not clear from our data, it could be that Mauritania is a society with engrained and culturally sanctioned norms and values that directly influencs decision-making. As these data were collected over a decade ago, information from other sources also provides useful insights into the problem of gavage in contemporary Mauritania. Indeed, there appears to be a movement toward an even more dangerous form of ''fattening'' young women to make them more desirable: the use of pharmaceutical drugs. A U.N. Population Fund official in Mauritania reported that: ''for the little girls that are being force-fed, the practice is getting more dangerous; before they used camel milk, nowadays the girls are force-fed with chemicals used to fatten animals'' [25,39]. Similarly, a 2011 State Department report begins with encouraging news of a decrease in gavage, but ends with the same troubling mention of chemical substances being used to achieve a fattened state [40]. Despite the legislation banning harm to children, both FGM and gavage persist. Importantly, no one has ever been prosecuted for FGM or gavage in Mauritania. A prominent child rights lawyer in Mauritania is quoted as stating: ''I have never managed to bring a case in defense of a force-fed child, as the politicians are scared of questioning their own traditions'' [7]. A comprehensive strategy to combat FGM and gavage should adopt a community-led developmental-based approach to include a plan to empower women economically and politically, especially in rural areas, and a plan to reduce illiteracy. Such a strategy cannot be implemented without substantial involvement and participation of religious leaders, community leaders, women and both male and female youth. Religious leaders could convey the message that FGM and gavage are not mandated by religion; public health and medical professionals could explain the sexual psychological and reproductive health complications associated with the practices; women and youth associations could focus on discussing and disseminating correct information related to these practices; and the government must focus on protecting the rights and welfare of children through the enforcement of law. National policies and legislations may not change cultural behavior, but such mechanisms could create a framework for openly discussing the public health and human rights concerns of such practices. This study has a few limitations worth mentioning. First, the analyses performed were done using cross-sectional data, thus precluding any causal inference. Second, the data were collected in 2000; although unlikely, given information from other sources, it is possible that the current situation related to FGM and gavage is different from 10 years ago. However, the recent DHS conducted in 2010 did not include gavage, making the data used in this study the only nationally representative survey data available that contains both practices. Lastly, given the sensitive nature of both FGM and gavage in the cultural and legal sense, these data may be prone to recall bias and bias related to self-reported activity. In conclusion, this study lends insight into two culturally sanctioned practices that have implications for the protection of children. These results should serve as a basis for the development of targeted interventions and strategies to change cultural perceptions. Until gavage is explicitly stated to be illegal, and current FGM laws criminalizing the act are enforced, reducing these practices will be a difficult task and countless children will continue to be subjected to these cruel and harmful practices. --- Author Contributions Conceived and designed the experiments: NO JK JB JR. Analyzed the data: NO JK JR. Wrote the paper: NO JK JB JR.

Objectives: To establish the prevalence of female genital mutilation (FGM) and force feeding (gavage) practices among children in Mauritania; to investigate factors related to FGM and gavage practices and attitude in Mauritania; and to explore implications related to the protection of children's rights and welfare. Methods: Data from the Mauritania 2000-2001 DHS were used in this analysis. Data were collected from men and women about their attitude toward the continuation of FGM and gavage; women only were asked if they ever experienced one of these practices. Chi-square statistics were used to investigate differences in attitude and practice of FGM and gavage by demographic characteristics. Binary logistic regression was used to identify socio-demographic factors related to FGM and gavage outcomes.The overall prevalence of FGM was 77% but varied depending on ethnicity. The majority of both female and male respondents favored the continuation of the practice (64% and 70%, respectively). Almost a quarter (23%) of women reported being force fed as a child and 32% of women and 29% of men approved the continuation of the practice. Gavage is almost exclusively practiced among Arabs.The practice of both FGM and gavage is ongoing, although the prevalence and attitude towards both appears to vary as a function of ethnicity, wealth, education, marital status, and age. Contextually relevant intervention and enforcement strategies are needed to challenge these cultural norms and protect the rights and welfare of children in Mauritania.

Introduction Medical science and technology advancements have improved quality of life over the years, leading to an increased life span [1]. As a result, people today are more likely than ever before to live longer and see their third and even fourth generations. This demographic shift not only increases the proportion of older individuals but also changes family structures and caregiving practices [2,3]. In China, where the older adult population constitutes about 17.8 percent [3], intergenerational support, as influenced by Confucian ideology, means that grandparents often play the role of caregivers to their grandchildren. While several social and economic implications have been attributed to this [4][5][6], a particular area of interest is how the arrangement tends to impact the life satisfaction of older adults in the context of active aging. Active aging, which emphasizes maintaining physical, mental, and social activity as individuals age, constitutes a central policy framework for coping with aging across countries worldwide [7]. Likewise, it is becoming a national priority in China as the population ages. The Chinese government's "Opinions on Strengthening Work for the Older People in the New Era" released in November 2021, called for integrating positive aging concepts and healthy aging into the economic and social development process [8]. This strategy aims to enhance older adults' health, social participation, security, and improvement of quality of life as they age. Quality of life is often measured by perceptions about life satisfaction [9]. As China continues to address population aging, improving the life satisfaction of older people has garnered attention from academic and social circles [10]. This arises from recognizing that various factors affecting older individual's life satisfaction and overall well-being can foster healthy aging, alleviate the strain on social security and familial caregiving, and contribute to both family development and social harmony [11]. Grandchild caregiving involves actively investing time, energy, emotional, economic, and material resources in caring for underage grandchildren [12,13]. Caring for grandchildren is the most common life event in later life and may either constitute a source of satisfaction or stress for older people [4,14]. Hence, exploring how caring for grandchildren affects older people's life satisfaction in the context of an aging population and active aging is crucial. In Chinese culture, it is common for grandparents to assume the responsibility of caring for their grandchildren, which can effectively alleviate the anxiety associated with parenting and the burden of childrearing [5,15]. This arrangement can also ease the financial strain on adult children [6]. Despite a significant decrease in families with multiple generations, the proportion of three-generation families (grandparents with adult children and grandchildren) remains significant, highlighting the crucial role of older people in caregiving, families, and society [16]. While previous research underscores the advantage of grandchild care, there is a growing concern regarding its potential adverse effect on the wellbeing of grandparents [17]. The current body of domestic and international research lacks a consensus regarding the influence of grandchild care on the life satisfaction of older individuals. Additionally, there is a dearth of systematic examinations of the underlying mechanisms that contribute to this impact. This paper employs data from a comprehensive social survey conducted in China to investigate the association between different forms of caregiving and the life satisfaction of older individuals. Considering the mediating effects of individual characteristics such as aging attitudes and situational factors such as support in grandchild care, the paper investigates the impact of grandparenting on life satisfaction among older Chinese adults. --- Literature review --- Role strain theory and role enhancement theory Role Strain Theory and Role Enhancement Theory are relevant to exploring relationships between caring for grandchildren and the life satisfaction of older people [17]. The role strain theory posits that when individuals occupy multiple social roles or face conflicting demands within a particular role, they may experience stress, anxiety, and a decreased sense of well-being [18]. These roles may include those related to work, family, community, or other aspects of an individual's life. Based on this theory, older people caring for grandchildren may experience pressure due to conflicting roles, which can become a burden and a source of stress. This can negatively affect their health and well-being, ultimately reducing their overall life satisfaction. In contrast, Role Enhancement Theory suggests that individuals who successfully fulfill their social roles may experience greater happiness, satisfaction, and wellbeing. According to this theory, meeting the expectations and demands of one's social roles can provide individuals with a sense of purpose, meaning, and fulfillment in life. Additionally, the theory proposes that the more roles an individual has, the more opportunities they have for social interaction, personal growth, and achievement [19]. Based on the theory's propositions, when older people take care of their grandchildren, it not only adds to their multiple role experiences but also enhances their sense of value. Additionally, it compensates for the reduction of social networks that often comes with old age, leading to improved social integration and, ultimately, a positive impact on their well-being. --- The effect of caring for grandchildren on life satisfaction among older people Intergenerational care is a common phenomenon worldwide [20]. Domestically and internationally, researchers have started to focus on the effect of grandparenting on the well-being of older adults, but the findings are mixed [21]. Some studies have reported that caring for grandchildren increases the physical burden [22] and emotional stress [20] in older individuals, resulting in cognitive decline [23]. Additionally, studies have also suggested that grandparents who care for their grandchildren may be unable to participate in other social activities, which lowers their entertainment and leisure time, potentially increasing their risk of depression and decreasing their life satisfaction [24]. On the other hand, some scholars argue that grandparenting can improve the physical and mental health of older adults [25,26], increase their happiness [27], and extend their life span [28]. A study of 20 European countries found that caring for grandchildren significantly increased the subjective well-being of older adults, regardless of the care intensity [29]. However, studies suggest that the protective effects of caring for grandchildren vary by gender, as older women seem to benefit more than men [30]. --- Intergenerational Support as a Mediator Research on the relationship between grandparenting and life satisfaction in China has explored various mediators, including social support. Chinese cultural norms often dictate strong intergenerational relationships, where older individuals who care for their grandchildren tend to receive increased care and support from their own children [31]. This support can have a profound impact, mitigating the negative effects of stressors related to grandchild care [31,32]. Empirical evidence has shown that this support can enhance cognitive function, reduce depressive symptoms, and ultimately increase life satisfaction among older grandparents [33,34]. A recent study that utilized longitudinal data involving 9,219 adults over 45 years [35] showed that caring for grandchildren positively correlates with better mental well-being among Chinese grandparents. Portions of the findings further revealed that children's support significantly mediates the impact of parenting experience on grandparents' mental well-being. However, despite the existing studies on the mediating role of social support, a literature gap still needs to be addressed. While the literature has recognized the positive effects of intergenerational support, it's important to delve deeper into the specific dimensions of this support (economic, instrumental, emotional) and their differential impact on life satisfaction. --- Aging attitude as a mediator Aging attitude represents a critical psychosocial factor that can significantly influence an individual's wellbeing and life satisfaction. Aging attitude refers to an individual's perception of the aging process, and it can be influenced by personal experience and social interaction, including caring for grandchildren [17]. Although a well-established association exists between aging attitudes and well-being/life satisfaction among older adults in China [11,36], the specific mediating role of aging attitudes between grandparenting and life satisfaction remains understudied with limited evidence. The role of aging attitudes in mediating the relationship between grandparenting and life satisfaction is vital in Chinese culture, where grandparenting often serves as a means of social integration and expansion of social networks for older individuals [37]. Older individuals who view grandparenting as an opportunity for personal growth, fulfillment, and social connection are more likely to have positive aging attitudes [19]. These positive attitudes may contribute to increased life satisfaction. Conversely, those with negative aging attitudes may perceive grandparenting as a burden or a source of stress, potentially decreasing life satisfaction [20]. In summary, the inclusion of aging attitude and intergenerational support as mediating variables in the relationship between caring for grandchildren and life satisfaction among older Chinese individuals is justified by the need to. These mechanisms not only align with theoretical frameworks related to well-being and aging but also have empirical support from previous studies. By exploring the specific dimensions of intergenerational support and the nuances of aging attitudes, this research aims to provide a deeper and more nuanced understanding of the intricate dynamics involved in how grandparenting impacts the overall life satisfaction of older adults in the Chinese context. --- Data, variables, and methods --- Data This study analyzed the 2014 Chinese Longitudinal Aging Social Survey (CLASS) data. The CLASS is a large social survey conducted by Renmin University of China, aimed at collecting data on the social and economic conditions, health, family status, and social support of older people aged 60 and above. The project aimed to gain a comprehensive understanding of the various problems and challenges that older adults face during the aging process and provide an important theoretical and factual basis for China to address and solve the problems related to aging. For the purpose of this study, older adults without minor (underage) grandchildren and those with missing key variables such as aging attitude and intergenerational support were excluded from the sample. The final sample size included for analysis was 5363. --- Variables --- Dependent variable Life satisfaction is an individual's subjective assessment of their quality of life [36]. The CLASS questionnaire measures an individual's life satisfaction with the item "Are you satisfied with your current life?" (Very dissatisfied = 1, relatively dissatisfied = 2, generally = 3, relatively satisfied = 4, very satisfied = 5). This article operationalized this variable as a continuous measure. --- Independent variable The study measured the independent variable of grandchild care by asking older people about the time spent caring for their minor grandchildren in the past 12 months. The questionnaire had six response options ranging from every day to rarely or never. The paper considered older adults to be caring for their grandchildren if they cared for at least one minor child. A dummy variable was created with a value of 1 for those who reported caring for their grandchildren and 0 for those who did not. The CLASS questionnaire included questions about five children of the respondents. --- Mediating variables This paper considers two mediating variables: aging attitude, which is further divided into self-aging attitude and general aging attitude, and intergenerational support, which is further divided into intergenerational economic support, intergenerational instrumental support, and intergenerational emotional support. The 2014 CLASS questionnaire included seven items from the Attitudes to Aging Questionnaire (AAQ), with a 5-scale rating ranging from "totally disagree" to "fully agree." The titles of these items are as follows: (1) I feel like I am getting old; (2) In my opinion, growing old means losing out on many things; (3) As I age, it becomes harder to make new friends; (4) I feel excluded because of my age; (5) The older one gets, the better they are able to handle life's problems; (6) Wisdom comes with age; (7) There are many enjoyable things about growing old. In this paper, the scores of the first four questions are summed to create a score for "self-aging attitude, " and the scores of the last three questions are summed to create a score for "general aging attitude. " The reliability coefficient for "self-aging attitude" is 0.70, while "general aging attitude" is 0.61. In this study, the support received through caring for grandchildren was categorized into three dimensions: intergenerational economic support, intergenerational instrumental support, and intergenerational emotional support. Each dimension was assessed using specific topics and response options, as outlined below: 1. Intergenerational economic support: Participants were asked if their child had ever given them money, food, or gifts, and also what it was worth. Response options ranged from Not given = 1; 1-99 yuan = 2; 100-499 yuan = 3; 500-999 yuan = 4; 1000-1999 yuan = 5; 2000-3999 yuan = 6; 4000-6999 yuan = 7; 7000-11999 yuan = 8; and 12,000 yuan or above = 9). 2. Intergenerational instrumental support: Participants were asked to indicate how frequently their child assisted them with household chores in the past 12 months. Response options are as follows: Almost not = 0, several times a year = 1, at least once a month = 2, at least once a week = 3, almost daily = 4). 3. Intergenerational emotional support: Participants were asked to provide a rated response to the question, "Do you think this child does not care enough about you"? Never = 1, occasionally = 2, sometimes = 3, often = 4). --- Control variables This research draws on the literature to explore two broad classes of confounding factors [32]: sociodemographic characteristics and health status and residential factors. The first category includes gender, age, marital status, education level, urban or rural residency, personal income, and work status. Marital status is categorized as widowed, divorced, or unmarried (0) and married with a spouse (1). Education level is classified as primary school or below (0) and junior high school or above (1). Household registration attributes are categorized as rural (0) or urban (1). The second category of variables includes health status variables, such as chronic diseases and instrumental mobility of older people, and living pattern variables, such as whether they live alone (living alone = 1; living with others = 0). --- Statistical analysis The analysis in this paper consists of several steps. First, a multiple linear regression model (OLS) is used to examine the influence of caring for grandchildren on the life satisfaction of the older people. Second, the propensity score matching method (PSM) is employed to test the stability and causality of the relationship between caring for grandchildren and life satisfaction. Finally, mediation analysis is conducted to analyze the mediating roles of aging attitude and intergenerational support in the association between grandchild caregiving and life satisfaction. All analyses were conducted using STATA 12.0 software. --- Results --- Descriptive statistics --- OLS result Table 2 shows OLS regression results of the life satisfaction of older people from the entire sample. Model 1 includes control and independent variables, while Model 2 adds aging attitude (self-aging and general aging) and intergenerational support (economic, instrumental, and emotional). Model 1 indicates that caring for grandchildren is significantly and positively associated with life satisfaction (<unk> = 0.060, p <unk> 0.05), and the model explains 5.0% of the variance. However, Model 2 shows that self-aging attitude, general aging attitude, and intergenerational support are significantly and positively associated with life satisfaction while caring for grandchildren is not. The model explains 8.4% of the variance. These OLS results suggest that aging attitudes and intergenerational support may mediate the relationship between caring for grandchildren and the life satisfaction of older adults. The model has no multicollinearity (VIF <unk> 2 in both models). --- PSM results This study employed propensity score matching (PSM) to confirm the correlation between caring for grandchildren and the life satisfaction of older people. Three PSM methods were utilized: nearest neighbor, caliper, and kernel matchings. The PSM results (Table 3) showed a good matching effect, and the average treatment effect (ATT) of caring for grandchildren on the life satisfaction of older people was around 0.07-0.08, indicating a significant impact. These results were consistent with the OLS regression outcomes, suggesting that caring for grandchildren was a vital predictor of the life satisfaction of older people. Therefore, it can be concluded that the effect of caring for grandchildren on the life satisfaction of older people was robust. The analysis of density distribution maps was conducted on both the control and treatment groups before and after employing various propensity score matching (PSM) methods to assess the effectiveness of matching. The results demonstrated a notable reduction in the disparity between the post-treatment and control groups following the matching process, indicating successful matching. Figure 1 visually represents the density comparison before and after matching, utilizing the kernel matching method. --- Mediation analysis This study examines the mediating effect of aging attitudes and intergenerational support in the full sample. The results of the mediation effect tests are presented in Table 4. The table shows that by including aging attitude and intergenerational support variables, the direct effect of caring for grandchildren on the life satisfaction of older people became insignificant (<unk> = 0.020, P = 0.453), while the mediation variables showed a significant mediating role (<unk> = 0.040, P = 0.000). Aging attitudes and intergenerational supports completely mediate the relationship between grandchild care and life satisfaction among older adults. The direct effects of caring for grandchildren on selfaging attitude (<unk> = 0.626, P = 0.000), general aging attitude (<unk> = 0.363, P = 0.000), economic support (<unk> = 0.114, P = 0.020), and instrumental support (<unk> = 0.574, P = 0.000) were significant. However, the direct effect on emotional support was not significant (<unk> = -0.002, P = 0.953). The direct effects of self-aging attitude (<unk> = 0.014, P = 0.000), general aging attitude (<unk> = 0.030, P = 0.000), economic support (<unk> = 0.036, P = 0.000), instrumental support (<unk> = 0.027, P = 0.006), and emotional support (<unk> = 0.100, P = 0.000) on life satisfaction were significant. Figure 2 provides a visual representation of the specific mediation effects observed in the study. Path 1 illustrates that caring for grandchildren has a direct effect on self-aging attitude, and this effect mediates a small but significant increase in life satisfaction (mediation size = 0.009). Similarly, Path 2 demonstrates a direct effect of caring for grandchildren on general aging attitude, contributing to increased life satisfaction (mediation size = 0.011). Path 3 indicates that intergenerational economic support has a direct effect on life satisfaction, with a mediation size of 0.004. Path 4 reveals that instrumental support also has a direct effect on life satisfaction, with a mediation size of 0.015. However, path 5 does not show any significant mediation effect of emotional support. Overall, these mediation tests demonstrate that aging attitudes and intergenerational support can play a mediating role in the relationship between caring for grandchildren and life satisfaction. --- Discussion Using 2014 CLASS data, this study found that caring for grandchildren positively impacts older people's life satisfaction, even after controlling for various factors. The study's findings provide substantial evidence of the full mediating role of aging attitude and intergenerational support in the said association. The study's findings, in several parts, are consistent with previous studies, particularly as it concerns the direct association between grandparenting and life satisfaction [26], as well as the mediating role of intergenerational support [12]. However, the finding on the mediating role of aging attitude is unique to the current study and, therefore, contributes to the existing literature and enhances the understanding of the intricate dynamics involved in how caring for grandchildren impacts the overall life satisfaction of older adults. First, the result on the positive association between grandparenting and life satisfaction among older adults is consistent with Tanskanen's study conducted with samples of European older adults [26], and supports the role enhancement theory [12]. Caring for grandchildren is an important family participatory activity that reduces the burden on adult children, increases their opportunities, and promotes family happiness [38,39]. Grandchild care also provides older people with a sense of accomplishment and emotional communication, enhancing their multifaceted roles and a sense of purpose and meaning in life, which improves their life satisfaction. This effect was found to be robust in the PSM test results. The core aspect of the current study is the investigation of the mechanism of the association, and the study found that caring for grandchildren impacts the life satisfaction of older people through individual and situational factors defined as aging attitudes and intergenerational support. Specifically, grandchild care positively affects the aging attitudes of older people, which is consistent with previous research in the context of China [12]. This significant mediation effect may be attributed to two factors. First, aging is often associated with a shrinking network size [38], the responsibility of grandchild care, however, could provide an avenue for older adults to socialize more and also increase activities of daily living (ADL). Studies have shown a positive association between network size increased ADL and improved life satisfaction among older adults [39,40]. Secondly, as supported by the role enhancement theory, caring for grandchildren enables older people to realize their value to society and their families, promoting a more positive attitude towards aging and improving their life satisfaction [36]. Furthermore, the finding showed that economic and instrumental support mediates the relationship between grandchild care and life satisfaction among older Chinese adults. Grandparents involved in grandchild care can significantly relieve their adult children of the burden of childcare, allowing them more latitude to focus more on their careers or other personal goals [37]. As a reciprocity gesture, adult children may be inclined to provide more financial and instrumental support to grandparents, strengthening their relationships and contributing to their life satisfaction [41]. This mutual benefit model of Chinese families could play an important role in improving the life satisfaction of older people from the perspective of social exchange theory [38]. Comparatively, the mediated effect size of the general aging attitude was substantially higher than the self-aging attitude. The findings imply that general attitudes about aging are more influential than personal attitudes. It is possible that people are more likely to be influenced by a societal construct or cultural attitude about aging than by their own personal attitude [42]. Also, among the three dimensions of intergenerational support, the result indicated that instrumental support had the most significant mediated effect size. This suggests that instrumental support, which includes financial/material resources and emotional resources, is vital in how older adults perceive life satisfaction [39] even when inundated with the stress of grandparenting. The findings of this study have significant theoretical and practical implications. Theoretically, this study contributes to the existing literature by revealing the complete mediating role of aging attitudes and intergenerational support in the relationship between caring for grandchildren and the life satisfaction of older individuals, casting light on the complex dynamics at play. Notably, it highlights that society and cultural views on aging might be more influential than personal attitudes and that instrumental assistance, such as financial and emotional resources, is critical in molding older adults' perception of life satisfaction. Practically, these insights underscore the importance of recognizing the social value of older people in supplementing child care and promoting family development. Policymakers and institutions should provide social support to older individuals who care for their grandchildren, encouraging them to maximize family benefits and promote their own healthy aging. Furthermore, older people should recognize their value to their children's family development and cultivate a positive attitude towards aging to achieve healthy aging through social participation. In addition, society should establish a cultural atmosphere that fosters respect and love for older people while rejecting age discrimination. This study has limitations in several aspects. Firstly, while the 2014 CLASS is relatively old, it was still used for this analysis because the more recent survey waves did not incorporate data on AAQ, an important mediating variable in this study. Though we acknowledge the importance of using more recent data, however, we believe that the dataset remains representative and relevant to our research questions. By using the same dataset for all variables, we maintain data consistency and comparability throughout the analysis. Also, given that China is largely a conservative society, there is likely to be some temporal stability regarding sociocultural and familial dynamics. To this extent, using the 2014 data does not necessarily diminish the relevance of our findings. Secondly, the cross-sectional data used in this study do not establish causality even with the PSM method. In addition, the study lacks data on how the age difference of the care recipient affects the relationship. Furthermore, caregiving behaviors may affect grandparents' health behaviors, and other mechanisms may be worth exploring. Lastly, the study only examines the relationship between caring for grandchildren and life satisfaction without considering the intensity of care. Despite these limitations, this research makes an additional unique contribution to knowledge regarding the association between caring for grandchildren and life satisfaction. Notably, the study sheds light on the mediating role of the aging attitude, further enhancing our understanding of the complex dynamics involved. --- Conclusion The study utilized data from the 2014 CLASS to examine how caring for children impacts the life satisfaction of older Chinese adults and the mediating roles of aging attitudes and intergenerational support. The finding showed that caring for grandchildren had a positive impact on older people's life satisfaction. It demonstrated that caring for grandchildren is a vital activity for older people that helps them develop positive aging attitudes and strengthens intergenerational support, thereby improving their quality of life. Encouraging social activities and giving full play to the important value of older people will create an age-friendly society and promote the harmonious development of the aging society. While our study has some limitations, as already acknowledged, future studies are welcome to improve upon it by examining how the duration and intensity of care influence the relationship between grandchild care and life satisfaction. --- Availability of data and materials The datasets analyzed in the current study is a secondary and publicly accessible data available in the repository of the Institute of Gerontology and National Survey Research Center of Renmin University, China, via http:// class. ruc. edu. cn/ index. htm --- Authors' contributions Conception, design, material preparation, data collection, and analysis were performed by Xinfeng Cheng. The writing of the first draft was performed by Xinfeng Cheng and Tolulope Ariyo. The revision and editing of the manuscript were performed by Tolulope Ariyo. All authors read and approved the final manuscript. " --- Declarations Ethics approval and consent to participate This study was performed in line with the principles of the Declaration of Helsinki. The survey protocol was approved by the Ethics Committee of Renmin University of China. Informed consent was obtained from all individual participants included in the study. --- Consent for publication Not applicable. --- Competing interests The authors declare no competing interests. • fast, convenient online submission • thorough peer review by experienced researchers in your field --- • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year --- • At BMC, research is always in progress. --- Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research? Choose BMC and benefit from:? Choose BMC and benefit from: --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Objective This study examines the impact of caregiving on older people's life satisfaction, focusing on the role of caring for grandchildren. The study considers individual characteristics (aging attitudes) and situational factors (intergenerational support) and aims to identify the mediating roles of aging attitudes and intergenerational support in the relationship between caregiving and life satisfaction.The study analyzed data from the 2014 China Longitudinal Aging Social Survey (CLASS), focusing on 5363 grandparents who reported providing care for their grandchildren in the 12 months before the survey. Life satisfaction was subjectively measured. The data was analyzed using multiple linear regression, propensity score matching, and mediation analysis.The study found that grandparents who cared for their grandchildren have higher levels of life satisfaction. Self-aging attitudes, general aging attitudes, intergenerational economic support, intergenerational instrumental support, and intergenerational emotional support fully mediated the relationship between grandchild care and life satisfaction.This study demonstrates that caring for grandchildren is a vital activity for older people that helps them develop positive aging attitudes and strengthens intergenerational support, thereby improving their quality of life. Hence, the government, society, families, and communities should provide more social support to older adults caring for grandchildren. This would not only benefit the health of the older people themselves but also promote intergenerational harmony and family development.

Introduction Many countries in Europe have long humanitarian tradition of receiving and resettling forcibly displaced individuals [1] and should provide equitable healthcare services to an increasingly diverse population [2]. Responding to changes in demographics and attaining equity in health can be viewed as a public health investment. However, for many European countries this is hampered by the lack of reliable knowledge of the health status and health needs of forced migrants in the early phase of resettlement [3]. Without adequate information, many resettlement-countries are unable to assess whether services are accessible for forced migrants and if needs are efficiently met [2]. The utilization of healthcare services is a multidimensional process that combines need for, and access to care. In an optimal scenario, use of services should be proportional to ones need [4]. Even though access of and use of services are inter-related, they are distinct parts of the health delivery process where utilization presumes access [5]. Factors related to access to services have been conceptualized in many ways, and commonly includes aspects on both the provider side and the user side such as accessibility, affordability, availability and appropriateness [6]. Both access and use of healthcare services is hence influenced by context, meaning that even where entitlements are formally established and financial barriers are lifted, access and use are influenced by resources required for good health, such as social support, education, and health literacy. Likewise, one might argue that additional factors related to the migration experience affect the use of health care services for forced migrants given the risk of exposure to external factors such as persecution, food insecurity, and violence. Exposures that can shape the forced migrants' health profile and subsequently their need for care [7]. This, however, has scarcely been researched. The Syrian refugee crisis remains the largest displacement crisis in the world, with 5.6 million registered refugees seeking transient safety in neighboring countries [8]. While in transit, healthcare services are often characterized by high privatization, fragmented between many different providers, making access to care difficult and costly [9]. For undocumented migrants, economic barriers are further aggravated with fear of detention or deportation if seeking healthcare [9]. Upon arrival in Norway, refugees are invited to a general health assessment, and have the same rights and entitlements to services as the resettlement country population. The Norwegian health care system offers universal coverage with relatively small out-of-pocket expenses. The general practitioner (GP) serves as a gatekeeper to secondary care, regulating the access to specialist and hospital care [10]. Primary care services are thus patient-driven while influx into secondary care is managed by healthcare providers. Through the resettlement process, refugees move from one set of health risk factors to another and can face multiple additional healthcare challenges along their journey. Few studies have focused on this change of context and environment, how it affects subsequent use of health care and whether adverse conditions affecting health and the use of health care services pre-arrival persists post migration. Applying a longitudinal design allowing a trajectory perspective, our study aimed to: (a) describe patterns of health care service use in Lebanon and Norway, (b) identify pre-arrival sociodemographic and migration-related predictors of health care service use post migration and (c) identify post-arrival factors associated with health care service use in the resettlement country. --- Methods --- Study design, participants, and data collection This is a two-time points follow-up study which is part of the Changing Health and health care needs Along the Syrian Refugees' Trajectories to Norway (CHART) project [11], assessing health of Syrian refugees in Norway. Methods were carried out in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement guidelines and with national and European privacy legislation. In this paper, we focus on persons recognized as refugees by the United Nations High Commissioner for Refugees (UNHCR) accepted for third-country resettlement [1]. The methods have already been described elsewhere [12]. In brief, a baseline self-administered survey was conducted in Lebanon in 2017-2018, followed by a follow-up survey in Norway after one year. A total of 514 Syrian nationals from 16 and above attending the mandatory Norwegian Cultural Orientation Programme (NORCO) in the given period were included in the study in Lebanon. The Arabic baseline questionnaire was distributed during course time under the guidance of crossculturally responsive bilingual trainers. Follow-up measurements post-arrival were gathered through structured telephone interviews in Arabic. A total of 506 eligible subjects completed the baseline survey (98 %), of which 464 (92 %) were confirmed resettled in Norway and 353 completed the second questionnaire (70 %) (Supplementary Fig. 1). --- Dependent variables The main outcomes for this study are the utilization of a GP, emergency care (EC), outpatient and/or specialist care as well as hospitalization during the previous 12 months. These four main outcomes were assessed through the following questions: 'During the last 12 months, have you visited any of the following: a general practitioner, emergency care, outpatient care, specialist care (yes/no)' and 'Have you been admitted to the hospital the last 12 months? (yes/no)'. Given similarities in outpatient and specialist care in Norway, where the main point is to be assessed by a medical specialist, these two variables were merged into one. The two items are based on questions from The Nord-Tr<unk>ndelag Health Study (HUNT) [13]. --- Independent variables Self-rated health (SRH) was measured as an indicator of the need for healthcare at both time points. We used a validated single-item question: "How do you consider your health at the moment?", with a five-point Likert scale ranging from very poor to very good. The item was dichotomized merging'very poor' and 'poor' indicating poor SRH versus non-poor SRH. The SRH-item has shown acceptable validity and reliability among Arabic speakers and in refugee populations [14,15]. Additionally, we measured quality of life (QoL) using the WHO Quality of Life Scale (WHOQOL-BREF), a transcultural instrument previously validated in Arabic [16], which includes a total of 26 questions on physical health, mental health, social relationships, and environment [17]. Each item is rated on a five-point Likert scale with a higher score indicating a better QoL. Raw scores were transformed creating domain scores within the range of 4-20 by multiplying the average of the items in each domain by four, in accordance with the user's manual [17]. Perceived social support was measured with the 7-item ENRICHD Social Support Instrument (ESSI) [18]. A total score is the sum of all items with higher scores indicating better social support. A binary measure for high social support defined as having answered > 2 on at least two items and a total score of > 18 was created, based on the definition of low-social support [18]. ESSI has previously been validated among Syrian refugees [19]. Sociodemographic variables encompassed age, gender, primary language spoken, marital status and level of education. In addition, we assessed Health Literacy through the single-item literacy screener (SILS): "How often do you need help reading written material from your doctor or pharmacy?" With a five-point Likert scale. Scores higher than 2 point to difficulties with reading health-related material. We also inquired on migration-related factors such as time since the flight from Syria, migrating alone or with family, residence permit in Lebanon, and possible exposure to traumatic events with The Single General Trauma Item [20]. The entire questionnaire was in Arabic; it contained questions already translated and validated and those sections that were not went through a standardized translation process [21]. --- Statistical analysis We present sociodemographic and migration-related characteristics as counts and proportions for categorical variables, medians, and interquartile ranges (IQR), and means and standard deviations (SD) for continuous variables (Table 1). Selection bias between the cohort and the loss-to-follow-up group was assessed using <unk> 2 -statistics and independent group's t-tests (Supplementary Table 1). We used a Sankey chart to visualize the changes in use of services before arrival and after resettlement by creating trajectory variables with the proportions going from use to no use and vice versa or no change in outcomes (Fig. 1). Changes in the use of health services from baseline to follow-up were also analyzed using generalized estimating equations (GEE) with data in long format with two observations per individual and "wave" as a binary covariate (Table 2). We applied a log-link and binomial distribution and reported exponentiated regression coefficients as risk ratios (RR) with 95 % CI. We used multivariate analysis to evaluate factors associated with the use of healthcare services in Norway looking at selected sociodemographic and migrationrelated factors as well as self-perceived health status and QoL at baseline and follow-up. First, we looked at baseline characteristics in Lebanon as predictors for the use of services after arrival in Norway. Thereafter, we looked at characteristics while in Norway and associations with the use of services in Norway. We used log-binomial regression analysis reported as risk ratios with 95 % confidence intervals in two models; (1) unadjusted (2) adjusted for potential confounders for the total effect of each characteristics on the outcome based on results from a directed acyclic graph (DAG) depicted in supplementary Fig. 2. The DAG was constructed using the software DAGitty [22]. For instance, for the total effect of health literacy at baseline on use of health services in Norway, age, gender, and education were potential confounders, while SRH at baseline was considered as a mediator and not adjusted for. In cases where convergence was not achieved in log-binomial regression analysis, Poisson regression was used with robust error variance (Table 3) [23]. Missing values were handled through listwise deletions. An alpha value of 0.05 was considered statistically significant. We analyzed the data using STATA/IC software, version 16.0, (StataCorp LLC, Texas, USA). --- Results --- Characteristics of the study population Sociodemographic characteristics and self-perceived health and QoL of this cohort has been published elsewhere but are stated in Table 1 for the sake of clarity. We included data from 353 participants in the final analysis (supplementary Fig. 1). The respondents did not differ from the loss-to-follow-up group in terms of age or gender but had higher health literacy (supplementary Table 1). --- Use of healthcare services and changes in use from Lebanon to Norway Of the 353 participants, 33 % visited a GP in Lebanon, 32 % visited outpatient/specialist care, 16 % were hospitalized, and 10 % used EC in the 12 preceding months at baseline (Table 2). In Norway, the use of a GP increased to 85 % and the use of EC to 18 % while hospitalizations remained the same and outpatient/specialist care visits dropped to 16 %. In Fig. 1, we present Sankey charts showing trajectories of healthcare service use. Most participants did not use EC, outpatient/specialist care, or hospital care neither at baseline nor at follow-up. There were 16 % new reports of EC use at follow-up, while 9 % used this in Lebanon but not in Norway. For specialist/ outpatient care, 10 % reported new use while 26 % reported using this in Lebanon but not in Norway. The biggest change in trajectory is the increase in the use of GP from pre-arrival to after resettlement with 58 % new reports of use. --- Pre-arrival predictors of use of health care services in Norway Increasing age was significantly associated with the use of EC services and hospitalization after arrival (Table 3). No other significant associations between pre-arrival sociodemographic factors and the use of healthcare services at follow-up were found. With regards to health status pre-arrival, we found that poor SRH was significantly associated with increased risk of EC use after arrival, while lower scores in the social relationships' domain of QoL (i.e., poorer social relationships) were significantly associated with use of EC after arrival. With regards to migration-related factors, not having a residence permit in the transit country was significantly associated with the use of EC after arrival. After-arrival factors associated with the use of health care services in Norway When in Norway, increasing age was still significantly associated with use of EC services and hospitalization (Table 3). Likewise, increased health literacy was significantly associated with use of GP, EC and hospitalization. Similarly, high social support (ESSI) was significantly associated with increased risk of EC use, use of outpatient/specialist care and hospitalization, and increasing education level was associated with hospitalization. When looking at health status, we found that poor SRH was significantly associated with the use of both EC and hospitalizations. Generally, lower scores in the different QoL dimensions were associated with higher use of services. However, higher scores in the environmental domain of QoL were significantly associated with use of a GP. --- Discussion This study provides data on health care utilization before and after resettlement assessed at two different locations and time points following the journeys of the same participants and therefore incorporates factors from the pre-arrival context as possible predictors for later use. We find that not having a residence permit and having poor health status pre-arrival predict the use of services after resettlement. Poor SRH was significantly associated with use of services both in Lebanon and in Norway suggesting a stable association along the migration path. For the post migration stage, we find a significant association between the use of healthcare services and increasing health literacy (SILS), high social support (ESSI), education and poor QoL. These factors did not seem to influence future health care behavior while in transit, suggesting phenomena subjected to change with time and context. Also, we find an increase in GP and EC use after resettlement and a decrease in outpatient/specialist care while hospitalization rates do not change pre-and post-resettlement, probably mirroring the health care system in the country of stay at each period. Use of GP services more than doubled pre-and postresettlement. This rate (85 %) is slightly higher than that of the resettlement country population in Norway, where 75 % reported use of GP in the last 12 months in population-based data [24]. Comparing numbers between surveys is encumbered with uncertainties, but we believe some of the differences in GP utilization between our sample and the resettlement country population can be explained by the fact that in some Norwegian municipalities, the general health assessment upon arrival is performed by a GP. Despite having a separate question for the general health assessment, we assume some participants might have had difficulties distinguishing between the two alternatives as both entails contact with a primary care doctor. Another possibility is that some refugees were derived to a second visit by the GP at the first encounter for their general health assessment. In any case, it is important to acknowledge the key opportunity GPs have in responding to the need of the refugee patient in early resettlement as the first point of contact. Previous studies have argued that refugee primary care services might reduce unnecessary EC use [25], showing that refugees who receive a health assessment shortly after arrival will be less likely to have an acute care visit in this period [26]. Furthermore, we found an increase in EC use from 10 % before arrival to 16 % after resettlement, which is similar to the utilization rates of the resettlement country population [27]. On the other hand, the use of outpatient/specialist care dropped from 32 % in Lebanon to 16 % in Norway. This decrease might be explained at the system level, since outpatient/specialist care services in Norway require a referral, usually from a GP, while other routes are available to access such care in Lebanon given a highly privatized health sector. When comparing with population-based data from the Norwegian population, 36 % reported having had contact with outpatient/specialist care the last 12 months [24]. A number twice as high as that of our population, but not adjusted for morbidity, so potential under-or overuse is not possible to determine with certainty. Furthermore, some of our respondents might have been referred by their GPs to secondary care, but still waiting for their appointments with a specialist at the time of the follow-up survey. However, the doctor-patient interaction is key in identifying patients needing a referral [28]. Previous research has shown that not speaking the same language is associated with decreased symptom reporting, fewer referrals to specialist care [29] and shorter consultation time [30], which also could explain our results. An inverse socioeconomic gradient in terms of utilization of outpatient/ specialist care has also been documented in Norway [31] that confirms privileged groups are those that avail most of services [32]. However, utilization of GP and hospital admissions, which is easier to access, was found to be equitable [33]. Similarly, a systematic review across Europe showed that outpatient visits for specialized care were generally used less often by migrants [34]. In our sample, hospital admissions did not change pre-and post-resettlement, which could point to hospital admissions having similar access thresholds across countries. Finding pre-migration predictors for use of health care in Norway can be of key importance to adequately prepare health services to the new migrant population. One novel finding in this study is that not having a residence permit in the transit country and having poor social relationships in transit was associated with higher use of emergency care the first year after resettlement. Generally, the lack of recognized documentation in a country complicates the availability of healthcare and one can assume that acute and/or chronic diseases left uncared for contribute to higher use of care post-resettlement. Hence, securing minimum acceptable living conditions for refugees in transit countries should be a priority concern globally. The strongest correlation we found was the one between poor SRH and health care utilization, signifying the concordance between need for care and use of care. Perceived poor health status seems to be a stable factor as it holds for both pre-arrival health status and after resettlement, even though the association after resettlement is stronger. While the association between health need and health care utilization is well-known [35], our study highlights the stability of this association along the migration trajectory. Post-migration, we found associations between use of services and higher health literacy, higher education, higher social support (ESSI), and low levels of QoL. High health literacy drops from 56 % in Lebanon to only 7 % in Norway, pointing to challenges with a new language and a different health care system, while high social support (ESSI) somewhat unexpectedly increases from 35 % in Lebanon to 60 % in Norway. This increase might be explained by the fact that most quota refugees are resettled as families and some are re-united with extended family members preceding them to the resettlement country. Easier access to online communication and established support networks upon arrival can also explain this increase. Why persons with high health literacy, higher education and high social support have increased probability of use while in Norway but not in transit is difficult to answer but we assume these factors become more important in a context where there is universal health coverage, and no economic barriers to health care. For the concept of social support and social relationships, we found associations pointing in opposite directions. Poor social relationships measured with WHOQOL-BREF while in Lebanon was associated with use of EC after arrival. When in Norway, high social support (ESSI) was associated with use of EC, outpatient/specialist care and hospitalizations. Likewise, poor social relationships (WHOQOL-BREF) were associated with outpatient/specialist care and hospitalizations. We believe some of this can be explained by measurement differences in social support instruments, not capturing the exact same phenomenon. The social relationships domain in WHOQOL-BREF as part of QoL only consists of three questions (satisfaction with relationships, satisfaction with support from friends and satisfaction with sexual relationships) and has the concept of satisfaction in it while ESSI consist of 7 questions and asks directly if you have someone available to talk to, receive advice, emotional support, receive help with daily chores etc. without assessing satisfaction. The environmental domain of QoL describes feeling of safety, satisfaction of living place, enough money to meet needs, and satisfaction with transportation. Interestingly, we found that higher scores in this domain were associated with use of a GP. This also confirms the inverse care law [32]. --- Strengths and limitations Working with a cohort with similar background arriving at the same time minimizing influence of contextual factors as well as a high response rate and the use of validated instruments add to the strengths of this study. However, certain limitations need to be considered when interpreting our data. We did not assess frequencies of contact with the healthcare services, only yes/no for use at least once. Because of this we are not able to separate between frequent users and persons who have only used the service once. This study has an explorative nature with a high number of statistical tests, which increases the risk of Type 1 error. We can therefore not rule out that some of the significant results are chance findings, especially those with p-values close to 0.05 (marked with one asterisk in Table 2). In addition, the variable health literacy is assessed with only one question (SILS) which is limited and has to our knowledge not been validated in a refugee population with poor language skills upon resettlement. Moreover, we deliberately changed mode of data collection from self-completion to structured interviews between the two time points which can introduce a possibility of interviewer bias, but in that way, we achieved a high response rate. Further, we should ideally have had a longer follow-up time to better assess changes with time. However, previous research has highlighted that we especially lack data on the first 5 years after resettlement [36]. Last, utilization of care is not equal with appropriate care or equality in quality of care, which we are unable to evaluate with the current study design. Despite these limitations, we believe our findings add important knowledge to the field of health services research for refugees, a group that is understudied in health system research. Based on our findings, we encourage resettlement countries to enhance primary care services in providing diversity-sensitive care given their role as first port of call. Possible under-use of specialist/ outpatient care among refugees and reasons for such differences warrants further research. People with undocumented status before arrival should be subjected to extra awareness to secure healthcare needs being effectively met at the primary care level. Social support and health literacy can be possible targets for future interventions to enhance accessibility of care. In conclusion, the use of healthcare for refugees clearly changes from the preand-post resettlement phase. Apart from entitlements and need, health care utilization is impacted by sociodemographic factors and migration-related factors. on previous versions of the manuscript. All authors (JHY, EMS, JI, EA, BK, WH and ED) read and approved the final manuscript. --- Availability of data and materials The datasets generated and/or analysed during the current study are not publicly available due to data protection regulations in Norway but are available from the corresponding author on reasonable request. --- Supplementary information The online version contains supplementary material available at https://doi. org/10.1186/s12913-021-06571-5. --- Additional file 1 Authors' contributions All authors (JHY, EMS, JI, EA, BK, WH and ED) contributed to the study conception and design. Material preparation, data collection and analysis were performed by JHY, EMS, JI and ED. The first draft of the manuscript was written by JHY and all authors (JHY, EMS, JI, EA, BK, WH and ED) commented --- Declarations --- Ethics approval and consent to participate We obtained ethical approval from the Regional Committee for Medical and Health Research Ethics of Norway (ref. no. 2017/377) and consent from the International Organization for Migration. Informed consent was obtained from all subjects. Confidentiality was ensured to all participants and data were de-identified and stored on a protected server. --- Consent for publication Not applicable. --- Competing interest The authors declare that they have no competing interests. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: Understanding the differential utilization of healthcare services is essential to address the public health challenges. Through the migration process, refugees move from one set of health risk factors to another and can face multiple healthcare challenges along their journey. Yet how these changing risk factors influence refugees' use of health care services is poorly understood. Methods: A longitudinal survey assessing health care utilization of 353 adult Syrian refugees was conducted; first in a transit setting in Lebanon and after one year of resettlement in Norway. The main outcomes are the utilization of general practitioner services, emergency care, outpatient and/or specialist care and hospitalization during the previous 12 months. Associations between use of healthcare services and several sociodemographic, migrationrelated and health status variables at both time points were found using regression analysis. We also analyzed longitudinal changes in utilization rates using generalized estimating equations. Results: The use of general practitioner and emergency care increased after resettlement while outpatient/specialist care markedly dropped, and hospitalization rates remained the same. Undocumented status and poor self-rated health (SRH) prior to resettlement were identified as predictors for use of health care after arrival. After resettlement, higher health literacy, higher education, higher social support and poor SRH and quality of life were significantly associated with use of healthcare services. Conclusions: Utilization of health services changes post migration to the destination country and are associated with migration-related and socio-demographic factors. Poor SRH is associated with use of services, both pre-arrival and post-resettlement. Our findings have implications for future resettlements, health care policies and service provision to newly arrived refugees with regard to both health needs as well as delivery of services.

INTRODUCTION Indonesia is a country with a diverse population consisting of various ethnicities, races, customs, classes, groups, religions, and social strata. The Indonesian society is a society with a highly complex level of diversity, a society with diversity (Afwadzi & Miski, 2021;Zakariyah et al., 2022). The richness and diversity of religion, ethnicity, and culture are like a double-edged sword. On one hand, this diversity is a treasure that must be preserved and defended because it has provided nuances and dynamics for the nation, and can become a wealth that may be rarely possessed by other countries. Muh Mawangir, Anica / Strengthening the Understanding of Religious Moderation at Islamic University However, on the other hand, the diversity of religion, race, and ethnicity can also become the starting point for disputes, vertical and horizontal conflicts (Nirwana & Darmadali, 2021;Idris & Putra, 2021). Looking at the potential of diversity above, in the era of open democracy, Indonesia certainly has a lot of differences in views and interests among its diverse citizens. It is important to manage these differences in a way that ensures that all aspirations can be channeled properly (Subchi et al., 2022;Husna & Thohir, 2020). Similarly, in matters of religion, our constitution guarantees the freedom of religious communities to embrace and practice their teachings in accordance with their beliefs and convictions. Our state ideology, Pancasila, places great emphasis on the creation of harmony among religious communities. Indonesia has even become an example for nations around the world in successfully managing its cultural and religious diversity, and is considered successful in harmoniously combining religious practice with statehood (Ministry of Religious Affairs of the Republic of Indonesia, 2019). Social conflicts and tensions on a small scale do indeed often occur, but we have always managed to overcome them and return to a consciousness of the importance of unity as a great nation, a nation blessed with diversity by the Creator. However, we must remain vigilant. One of the greatest threats that can divide us as a nation is religious-based conflict, especially when accompanied by acts of violence (Naim et al., 2022;Abidin & Murtadlo, 2020). The reason for this phenomenon is in the inherent predisposition of religion, regardless of its form or location, to exhibit a fundamental inclination towards prejudice, which is accompanied by intense emotional involvement and subjectivity. Consequently, it consistently engenders strong emotional attachments among its adherents. Religion is regarded as a hallowed and treasured entity, even by its most devoted followers. Excessive zeal for a particular interpretation of religious truth frequently results in hatred and conflicts, rather than leading to a serene and tranquil existence. This conflict, driven by religious beliefs, might affect distinct factions or denominations within the same religion, or arise between varied groups belonging to different religions. Usually, these religious conflicts begin when people blame each other for their different interpretations and understanding of religion. They feel morally superior and refuse to consider the perspectives and interpretations of others. In order to effectively handle the extensive range of religious beliefs in Indonesia, it is imperative to establish a clear vision and implement solutions that foster harmony and tranquilly in religious affairs. This can be achieved by placing emphasis on religious moderation, acknowledging and respecting various interpretations, and actively discouraging extremism, intolerance, and violence (Harahap et al., 2022;Idris et al., 2021). The essence of religious moderation lies in the pursuit of a balanced position between the two opposing ends of religious observance. One faction of the religious community staunchly adheres to a specific interpretation of religious texts, considering it as the only valid truth, and labelling other interpretations as abnormal. The group is usually known as ultra-conservative (Kamaludin & Purnama, 2021). Conversely, there exist religious adherents who exhibit an intense commitment to rationality, to the point of discarding the sacredness of their own religion or compromising their core convictions in order to display unwarranted acceptance towards adherents of different faiths. They are frequently identified as radical progressives. The moderation of both extremes is necessary, thus making the role of students pivotal in endeavours to enhance religious moderation education in society. According to Afif and Dwijayanto (2021), students have the ability to bring about significant transformations in the domains of education, society, and national affairs. The spirit of religious moderation needs to be strengthened, as there was evidence from field observations that students in higher education were susceptible to being infiltrated by radicalism, easily accepting fake news (hoaxes), arrogance, and fanaticism towards religious beliefs. Radicalism or crime in general is a natural condition of society, as the sociological reality shows that radicalism is found in almost all layers and forms of society, whether it is a simple or complex social structure (Supriyanto, 2022). Furthermore, the understanding of religious moderation as a solution to prevent negative radicalism is influenced by the internal condition of the individual itself and external factors from the environment. Internal conditions, such as psychological or mental health conditions, economic conditions, and so on, can affect this understanding. On the other hand, external or environmental factors can be influenced by the community, education, socialization, parenting, and so on. Among them, the researcher described that these young campus religious activists are also in the psychological phase of adolescence or late adolescence (Nugraha, 2020). At this stage, adolescents have started to organize their psyche and determine the direction of their lives with more direction. At this phase, the orientation shifts to external factors, with attention to issues surrounding their environment and society (Wasehudin & Syafei, 2021). During this time, adolescents become more critical, have an independent attitude, and thoughts. The process of searching for identity and self-organization is carried out in accordance with the perspective that is built through the knowledge and doctrine received from their group. Generally, these individuals feel like activists in the religious field, and thus, they must be able to act, behave, and conduct themselves in a religious manner that is filled with righteousness (Firmansyah, 2021). This can certainly awaken enthusiasm for religious knowledge that can support the attitude of personal righteousness for some individuals. The ideal Islamic behavior portrayed by the salafus shaleh is their role model. With the knowledge they can access, they try to display righteous behavior, including following the Sunnah in appearance and daily behavior. An interesting study for them is the popular jurisprudence (fikih) discussion about everyday Sunnahs. Generally, organizations among students, especially religious organizations, consider the veil phenomenon, which is said to be wearing a veil is not an obligation, but is seen as a more righteous attitude in practising religion. The attention of these campus religious activists is directed not only internally towards themselves but also towards their external world with criticism of their environment (Jailani & Suyadi, 2023). An ideal personality and an ideal environment are in accordance with the Islamic values that they believe in. In this understanding, a gap emerges between the idealized vision and the reality in their environment, which prompts them to criticize with the religious perspective they believe in. Generally, their attention is focused on piety and morality, which is why they view the external world through a moral lens. Due to this gap, moral panic arises, a feeling of discomfort and displeasure with behaviors deemed immoral according to Islam. The group experiences a moral panic when they feel that a social or cultural value they hold is threatened by a perceived social order disruptor, known as a 'folk devil.' The process of creating moral panic starts with the existence of something that is seen as a threat to social norms or the interests of the community; The media then portrays it in easily understandable symbols, which eventually create anxiety in society. As a result, responses emerge that ultimately drive social change in the community. In the context of the understanding of university students in a religious organization, they also view Western culture as a threat to Islamic morality, particularly towards the behaviors of young adults. One of their concerns is the issue of dating. Male-female relationships are highly restricted, therefore the phenomenon of dating among teenagers or college students is considered a bad thing (Wardi et al., 2023). One of the characteristics of the religious understanding of students (Abdullah & Nento, 2021), the characteristics can be described as follows; First, they are textualists. Students, in general, tend to follow the textual evidence as it is, what is said in the Qur'an and the Hadith of the Prophet, they tend to apply it as it is. Often, understanding the context of the verse or Hadith is not their concern. Similarly, interpretation receives less attention compared to translation. Secondly, the religious knowledge of students on campus is utilized to improve behavior in accordance with the Islamic teachings they understand. Therefore, their orientation towards understanding religion is aimed at building their own and their environment's morality. The existence of a gap or discrepancy between what they idealize and the behavior of society in general raises criticism towards behaviors that are deemed inconsistent with religious teachings. The National Counter-Terrorism Agency (BNPT) even asks the academic community to be vigilant of radical movements within the campus, especially towards new students as an entry point. The vulnerable group susceptible to being influenced and involved in radical movements are young people who are curious but whose personal conditions are not yet stable (Suyudi & Putra, 2022). In the perspective of sociology itself, radicalism or crime in general is a natural condition of society (crime is a natural part of society). This is said because the sociological reality shows that radicalism is found in almost all layers and forms of society, whether it be simple or socially complex communities. Thus, radicalism is divided into two types or forms. Firstly, individual radicalism (individual violence), which is radicalism carried out by an individual towards others. Secondly, group radicalism (group or collective violence), which is a form of radicalism carried out by a group of people towards an individual or another group (Henslin, 2016). In other words, the understanding of religious moderation as a solution to prevent negative radicalism is influenced by both internal individual conditions and external factors (environment). The continuing occurrences of terrorism and actions that lead to radicalism in Indonesia, particularly in the education sector, are concrete evidence of how the understanding and appreciation of the values of religious moderation are still low. Therefore, various approaches to counterterrorism and radicalism must always be pursued. The research on religious moderation has been conducted extensively to strengthen and realize attitudes of religious moderation amidst the flow of the times and the level of religious understanding of each layer of society. Firstly, Kawangung (2019) conducted a research examines the Religious Moderation Discourse in the Plurality of Social Harmony in Indonesia. The result found that religious moderation refers to achieving harmonious living by upholding human dignity. Secondly, Husna & Thohir (2020)'s research on Religious Moderation as a New Approach to Learning Islamic Religious Education in Schools reveals that religious moderation must continue to be upheld in schools to prevent extremism. Thirdly, Fawaid & Astutik (2023)'s research on efforts to build a religious moderation attitude through Islamic religious education among students in public universities. The findings show that the substance of the Islamic religious education curriculum should be directed by lecturers to strengthen moderate attitudes and provide guidance and mentoring to students. Fourth, a study conducted by Munir & Herianto (2020) explored the influence of mental health, organizational activity, and academic achievement on the level of understanding of religious moderation. The findings of the study show that these factors have an impact on students' understanding of religious moderation. And fifth, a study by (Zarkasyi, 2018) appraised the moderation of Indonesian Muslims with special reference to Muhammadiyah and Nahdlatul Ulama. The findings of this study show that Muslims in Indonesia, both Muhammadiyah and Nahdatul Ulama, have a long history of organization through educational institutions. Religious moderation can be realized, one of which is through their educational institutions. The literature does not provide information on how Islamic education, which is based on religious moderation, can enhance the comprehension of students, particularly those enrolled in the Islamic Education programme at the Faculty of Education and Teaching, Raden Fatah State Islamic University in Palembang. Thus, implementing deradicalization programmes at Islamic higher education institutions is one effective approach to enhancing students' comprehension of religious moderation. Hence, the author aims to investigate and explore this research thoroughly to provide research findings that would enhance the comprehension of religious moderation among university students, particularly in the Islamic Education programme at Raden Fatah State Islamic University Palembang. --- METHODS This study was a type of qualitative research with a descriptive analysis approach. The qualitative descriptive approach can be understood as research that describes qualitative data as it is or naturally. Then, the data is analyzed to reveal the facts that occur or appear (Sugiyono, 2014). This research involved 45 students of the Islamic Education program, Faculty of Education and Teaching Sciences, Raden Fatah State Islamic University Palembang, consisting of 25 students from the third semester of the 2020 academic year and 20 students from the fifth semester of the 2019 academic year. The sample for data analysis in this research was selected using purposive sampling technique. Data were collected through interviews and documentation, with a focus group discussion, in order to gain in-depth information. Primary data through interviews were conducted on December 16 and 17, 2021, in person. Each student was given questions related to their understanding of religious moderation and current religious issues such as news about religion in society, social media, fake news, radicalism, terrorism, fanaticism, and intolerance. From the questions asked, for secondary data documentation, the author gathered information in the form of books on religious moderation by the Ministry of Religious Affairs of the Republic of Indonesia in 2019, M. Quraish Shihab's book Wasathiyyah: Insights into Religious Moderation, and journals on religious moderation online. This data was employed as reinforcement, complements, and sharpens the data related to this research. --- FINDINGS AND DISCUSSION --- Islamic Education Islamic education aims to shape Indonesian individuals who have faith and piety towards the One Almighty God, possess noble character, and are capable of maintaining peace and harmony in both internal and interfaith relationships. Islamic education is also directed towards developing the abilities of students in understanding, internalizing, and practicing religious values that harmonize with the mastery of science, technology, and the arts. These messages are contained within Islamic teachings on religious moderation. Islamic education based on religious moderation will be more closely related to the methods that an educator will take and use in delivering and imparting learning materials on religious moderation. There are three learning strategies that can be taken by Islamic education based on religious moderation towards students: first, inserting religious moderation content into relevant course materials. In fact, some course materials already contain religious moderation content. The substance of religious moderation is already present in the learning curriculum at all levels and types of Islamic education within the Ministry of Religious Affairs environment, including Raden Fatah State Islamic University in Palembang. Secondly, optimizing learning approaches that can foster critical thinking, an appreciation for differences, respect for others' opinions, tolerance, democracy, the courage to express ideas, sportsmanship, and responsibility. Education, society, and national affairs are all areas where students, as Afif and Dwijayanto (2021) argue, can and should effect major change. This type of Islamic education approach based on religious moderation is carried out when transforming knowledge to students both inside and outside the classroom. Thirdly, organizing Islamic education programs through specific education, training, and preparation with a special theme on religious moderation. It can also be done by organizing religious courses with religious moderation content. Fourthly, reaching the evaluation aspect. Lecturers need to observe simultaneously to evaluate the achievement of the learning process that has been carried out with methods that can cultivate a moderate attitude. --- Religious Moderation From the research results obtained through student interviews and documentation, the researcher found awareness of the efforts made by students, both from 25 third-semester students of the 2020 cohort and 20 fifth-semester students of the 2019 cohort, regarding the importance of strengthening the understanding of religious moderation among students. The lecturers in the Islamic religious education program at the Faculty of Education and Teaching Science must make a strong and continuous effort to educate on the meaning of respecting differences of opinion among religious communities, including differences of opinion among students themselves. In lectures, both lecturers and students are required to understand various religious courses, both from the university curriculum and specialized courses in the study program, so that students are able to comprehend the importance of Muslims knowing and studying religious understanding, both within Islam and non-Islam, as expressed by a respondent (IN). As students, we must also be friendly and accepting of differences as long as they do not conflict with our beliefs as followers of a religion (interview with IN on December 16 th, 2021). Then the second, understanding the rules, provisions and signs in understanding Islam, the third is being careful in applying Islamic teachings that are not properly supported by understanding. Fourth, there is a strengthening of understanding of Islam in accordance with the correct guidance and fifth, being open to thought, and reform but still able to screen and be positive. From the outcome of the interview, it was emphasized that strengthening religious moderation must be positively charged and properly understand religion so that it is not misguided (interveiw with MD on December 17 th, 2021). Furthermore, this understanding is also corroborated by AD's opinion that he always tries to be open, understands every difference that exists, thinks wisely, and often travels or travels so that he can observe the outside world (interveiw with AD on December 17 th, 2021). Then, NP argued that it is our duty to accept differences but continue to unite in differences, do not let conflicts and prolonged problems arise and we must be in the middle position not to become a heater or stove (interveiw with NP on December 17 th, 2021). The Medina Charter, which was disseminated by Rasulullah ibn Abi Talib, is an answer to the social realities of the lives of religious people, as evidenced by the above opinion. This has been the case ever since the Prophet Muhammad PBUH first illustrated it. Bringing people together via shared diversity. On the intellectual, spiritual, social, and economic fronts. For this reason, it's important to look at the Islamic concept of religious moderation (wasathiyyah) and how it relates to the role of the state in implementing the ideals of justice and balance. A thorough investigation of the Prophetic traditions is necessary for the analysis and application of this idea. As part of this initiative, Muslims can look to the Medina Charter for guidance on how to conduct themselves in all aspects of their lives, not just their faith. This is why the following writer defines wasathiyaah (religious moderation) in Islam. Wasatha, also spelt Wasathiyyah, is the Quranic phrase for religious moderation. The five verses in the Qur'an where this word appears are: Q. S. Al-Baqarah 143, Q. S. Al-Baqarah 238; Q. S. Al-Maidah 89; Q. S. Al-Qalam 28; and Q. S. Al-A'adiyat 4-5 (Shihab, 2019: 4). Many Muslim scholars base their interpretations of the Qur'an on verse 143 of Surah al-Baqarah, which they describe as follows: "And likewise We have made you (Muslims) "the middle people" that you may bear witness to the people and that the Messenger (Muhammad) may bear witness to you. We have not made the Qiblah to which you used to turn, but that we may recognize those who follow the Messenger and those who turn back. Indeed, (the transfer of the Qiblah) is very hard, except for those who have been instructed by Allah. And Allah will not waste your faith. Indeed, Allah is the Most Merciful, the Most Merciful to mankind" The above verse is utilized by many experts as a starting point regarding religious moderation in this study, as described by Ar-Razi's opinion in M. Quraish Shihab (2020: 10) there are various meanings of the verse that have been delivered, first the word fair. This word is interpreted as the same attitude towards all directions that do not lead to one direction or ignore other directions. Secondly, the word wasath in the above verse means the best. Third, the most excellent or the best. The fourth meaning is if Muslims as ummatan wasathan means having a moderate or middle between excellence and deficiency in various matters. In addition, this opinion is also strengthened by M. Quraish Shihab that the nature of ummatan wasathan means that they can have differences in paths, all of which can be accommodated in a wide path as long as the path they take is characterized by peace. Allah hints that if they are truly different, but all can go hand in hand on one wide road, then they can lead to the pleasure of Allah SWT (Widodo, 2019;Muchtar et al., 2022). Thus, the mention of Muslims as ummatan wasathan is an effort so that they can display themselves as a chosen people who have attitudes full of justice, both during worship as individuals or during interactions as members of society. Islam offers teachings to maintain a moderate attitude (interveiw with WJS on December 16 th, 2021). This is also recognized by students of the Islamic education study program at the Faculty of Tarbiyah and Teacher Training, Raden Fatah State Islamic University Palembang, LA said that as the middle community and the people of the Prophet Muhammad, they must be an example for other groups (interveiw with LA on December 16 th, 2021). So that the word wasathiyah is located in the Muslim community needs to be positioned in the context of the community's relationship with other communities. Muslim individuals or communities can be regarded as witnesses (shahidan) when they are dedicated to moderation and human values (Manshur & Husni, 2020;Sofyan et al., 2022). Because, if the understanding of wasath refers to the context of moderation, then he demands Muslims as witnesses or as witnessed, to be used as a link for other people, and at the same time they are referring to the Prophet Muhammad PBUH as an example that they deserve to emulate as a witness to the justification of all their activities (interveiw with DNS on December 17 th, 2021). It can be said that the level of an individual's loyalty to moderation actually indicates the extent of their commitment to the value of justice. If an individual is more moderate and balanced, then their potential for acting with full justice becomes more open. Conversely, if someone is not moderate and tends to be biased, then they are more likely to be unjust. This idea shows the tradition of Islam, which Prophet Muhammad SAW directed his followers to often choose the middle path, which they believed to be the best path. Based on the hadith, Prophet Muhammad SAW advised the ummah to often choose the middle path, which they believe is the best path. In another hadith reported by Muslim, it was mentioned that the Prophet said, "The best of affairs is the middle course" (Muslim). Jabir Bin Samurah also once said, "I have prayed with the Prophet Muhammad PBUH many times and I found that his prayer and his sermon were always in the middle". In another narration, the word "wasath" is found when Ibn Abbas said that the Prophet Muhammad PBUH said, "O people, avoid excess exaggerated in religion, for those who came before you were destroyed by their excess in religion." (Mahrus et al., 2020;Mindani et al., 2022). After its definition was clarified, the concept of wasatiyyah represents the criteria and substance of a religious teaching that embodies the meanings of moderation, balance, justice, and the best option. These three explanations are interrelated because a moderate attitude often represents fairness and the best choice. Furthermore, the dimension of strengthening religious moderation among students needs to be understood as follows (Daheri, 2022): 1. Moderation in theology refers to the attitude of tauhid, meaning acknowledging and worshiping only one God. This attitude is positioned in the middle, as a balance between atheism, which is anti-theistic, and polytheism, or the worship of many gods. Tauhid means recognizing one God who is worthy of human worship. Tauhid also means believing that Allah is a single entity, without consisting of various elements or components. The concept of polytheism, which acknowledges many gods, is an irrational doctrine and falls under the category of ghuluw (exaggeration). 2. Moderation in worship, the Qur'an emphasizes that the creation of jinn and humans is intended for worshiping Allah SWT (Q.S az-Zariyat: 56). Worship means a submissive and obedient attitude towards Allah SWT, accompanied by a sense of admiration for Him. Some scholars mention that conceptually, all actions can bring about the pleasure of Allah SWT, with sincerity and pure intention from the conscience based on the principle enshrined in the Qur'an and the Prophet can be called worship. Categorically, the scholars divide worship into two categories: first, mahdlah worship (pure worship) which is a ritual individual worship, and secondly, social muamalah worship. Both are intended to build spiritual and social righteousness. 3. Moderation in the relationship between Islam and culture. Islam and culture should not be opposed to each other, as there are many good traditions and cultures, although they may not be found in a clear textual reference. The assumption is that Islam did not appear in a cultural vacuum. Islamic teachings have always been present in existing traditions, even though there may need to be a negotiation process of acculturation. Sometimes Islam will forbid certain cultural practices if they clearly conflict with the aqidah. At times, Islam changes and renews culture, when there is something that needs to be relevant to Islamic teachings, and there are times when Islam accommodates a culture that is in line with Islamic values or teachings. For example, the tradition of giving an engagement ring to the bride-to-be during a proposal, as is practiced in Egypt. Of course, this tradition does not have a clear hadith as evidence. However, in substance, it actually strengthens the engagement bond and serves as a sign that the woman has been proposed to. Therefore, for other men who want to propose to her, they will know that the woman has been proposed by someone else. This means that the tradition reinforces the sunnah of the Prophet which states, "It is not permissible for a man to propose to a woman who is already engaged to his brother" (H.R. Bukhari). 4. Moderation in treating women. Women's position, when referring to the perspective of the Qur'an, is very noble and honorable. Women are placed in a moderate position between belittlement and excessive exaltation. In general, women are equal to men. They are not placed under men, but they are also not placed above men. Furthermore, Abdul Mustaqim stated that regarding women, first of all, they are created equally with men. 5. Moderation in the relationship between Islam and the state. Islam gives the choice to Muslims to determine the system of government. It does not have to be an Islamic state formally. In principle, the state can represent a sense of security (baladan amina) (Q. S Ibrahim ayah 35-37) And prosperity (baldatun warabbun ghafur) (Q. S Saba' ayat 15). Moderation in the relationship between Islam and the state. Islam provides options for Muslims to determine the government system. It does not have to be an Islamic state formally. In principle, the state can represent a sense of security (baladan amina) and prosperity. Muslims will be free to make political ijtihad based on the preparation and social capital of the community. Indonesia deserves to be an example as one of the countries with the largest Muslim population that has implemented the concept of religious moderation in Islam. Indonesia is not a formally Islamic state and constitutionally, but not a secular state. Indonesia is a peaceful country that upholds high tolerance and harmony in society. However, Indonesia has a strong foundation, which is Pancasila. Pancasila is a philosophical basis built by the Founding Fathers during nation-building and governance. Pancasila is not a religion, and Pancasila does not have any conflicts with any religion. That Pancasila is a sincere effort of the founding fathers of Indonesia that has been agreed upon by all layers of the nation as the foundation of the state. All the principles contained in Pancasila are not in conflict with the teachings of the Quran and Sunnah. Pancasila can serve as an entity to preserve diversity. Therefore, if there is an effort to replace the foundation of the Indonesian state with the Islamic Caliphate system, it will result in a shift away from religious moderation in Islam in Indonesia. 6. Moderation in interacting with non-Muslims. Often, religion has two sides, gentle and harsh. The gentle side can be shown to the group for internal solidarity, while the harsh side is shown to the outside group as a defense mechanism. If these two sides are not guarded by practicing religious moderation, religion can trigger social problems. Therefore, tolerance is a way to avoid extremist and syncretic ideologies. Tolerance is not syncretism, which tends to standardize all religions. Tolerance is also not a proselytizing attitude that assumes there is no similarity between religions. From the dimension of strengthening the understanding of religious moderation above, we can see, reflect, and implement the true nature of religion. We agree that all religions have a strong commitment to carrying out their activities for their followers. However, all religions also agree to always teach their followers to avoid acts of violence, because no religion allows its followers to kill or harm others (Senata et al., 2021). All religions also teach their followers to be in solidarity, to respect, and to care for humanity in all aspects. The concept of actions guided by conscience is also taught by every religion because the goal of religion is for its followers to behave and speak according to the reality taught in the religion. Thus, in the end, all religions convey values of peace, equality, tolerance, and harmonization. In this regard, academic community, lecturers, and students continue to strive in efforts to strengthen the understanding of religious moderation in Islamic higher education institutions. --- Muh Mawangir, Anica / Strengthening the Understanding of Religious Moderation at Islamic University --- CONCLUSION The findings of the aforementioned study indicate that Islamic education at Raden Fatah State Islamic University Palembang employs a learning strategy centred around religious moderation. This approach aims to enhance students' comprehension and knowledge. This encompasses the integration of religious moderation content into the educational curriculum, enhancing learning methodologies, coordinating educational initiatives, providing students with specialised materials on religious moderation, and empowering instructors to evaluate students' comprehension of religious moderation. Moreover, it is imperative for the academic community, instructors, and students to collectively acknowledge and implement religious understanding in all aspects of life. This can be achieved by enhancing religious education through compulsory lectures in university courses, faculty courses, and study programme courses. Enhancing comprehension of religious moderation is crucial for students in the Islamic religious education programme. It guides them to effectively address and resolve religious matters within Indonesia's diverse culture. Based on this premise, all members of the academic community, including instructors and students, can achieve this goal by fostering camaraderie among religious communities, promoting peace, fostering a harmonious existence, advocating for equality and tolerance, adopting a moderate approach, preventing conflicts, avoiding selfishness, and fostering a sense of unity. Through collaborative efforts, it is possible to achieve longlasting religious moderation, while prioritising the preservation of unity over divergences. Moreover, students enrolled in the Islamic religious education programme at the Faculty of Education and Teaching Science set a precedent for students in other study programmes within the Raden Fatah State Islamic University Palembang community.

Religion may solve all human life problems, including violence, oppression, radicalism, terrorism, intolerance, fanaticism, and excessive behaviour in all areas of life, including education. Islamic higher education provides intellectual, social, and religious education to the community. Students, as the younger generation, lead civilization and show mature attitudes towards social issues, including rising religious issues and religious moderation. This study seeks to enhance the comprehension of religious moderation at Islamic University. This study employed a qualitative approach, utilising descriptive data analysis. Information was gathered through the means of interviews and documents. This study included a sample of 45 students enrolled in the Islamic Education program. The results suggested that implementing Islamic education centred on religious moderation to enhance students' comprehension of religious moderation in the Islamic Religious Education programme should be achieved through instructional approaches during lectures and mutual support from the academic community, professors, and students. Therefore, enhancing comprehension of religious moderation can foster a more profound understanding of adopting a peaceful, harmonious, non-anarchic, non-fanatical, non-arrogant, courteous, and respectful approach towards differing opinions while upholding the value of unity among students in the Islamic Religious Education programme at the University.